Description:A DEFECT INSPECTION SYSTEM AND AN ASSOCIATED METHOD

RELATED ART

[0001] Embodiments of the present specification relate generally to a defect inspection system, and more particularly to a system that accurately identifies defects in one or more rails of a railway track under dynamic ambient lighting conditions.
[0002] Rail tracks are subject to tremendous loads day and night to support transportation of goods and passengers. Continually subjecting the rail tracks to such loads leads to defects or anomalies associated with one or more rail track components, such as, rail joint, rail abrasion, fish plates and connection track component, switch gap, and track fastener. Track anomalies may affect operational conditions of a rail track, further leading to undesirable incidents such as derailing of coaches, halting and delay of transportation of goods. Rail tracks should, therefore, be periodically inspected for anomalies and to ensure safe operating conditions of the tracks and proactively performing any required maintenance to prevent accidents and to decrease the amount of time the rail track is under repair.
[0003] Conventionally, rail tracks are inspected manually by a track inspector who visually inspects the tracks by travelling along the tracks on foot or in a vehicle at a speed that allows the inspector to identify anomalies in the track structure. While manually inspecting tracks, minute defects, such as slightly increased gap between the switch rail and the main rail, may escape the eyes of the inspector. Manual inspection is also highly dependent on ambient lighting, and thus, can be performed only during the day owing to poor ambient lighting at night.
[0004] To overcome the drawbacks associated with manual inspection, certain automated methods are employed that use imaging systems including area scan cameras and fixed focus lenses to identify anomalies. Yet, these imaging systems are often unable to generate consistent, uniform intensity, and high contrast images due to the ambient light varying dynamically in the outdoor environment. For example, the imaging systems using area scan cameras and fixed focus lenses may fail to provide consistent output images when the track suddenly enters a tunnel from bright outdoor light or when a shadow is cast over a portion of the track. The images, thus captured under varying lighting conditions, are likely to be underexposed or overexposed including contrast and saturation issues. Further, when such an imaging system travels along the tracks, the distance or depth between the imaging system and the track varies dynamically as the imaging system may travel at different heights from the track. These height inconsistencies also cause a lack of focus in a fixed focus lens used in the image system, which in turn, causes the image system to capture defective and low contrast images. Current automated inspecting systems either reject such defective and low contrast images as inputs or identify a gap between the switch rail and the main rail inaccurately.
[0005] Certain present-day imaging systems employ line scan cameras in lieu of area scan cameras conventionally used in inspection systems to improve the low pixel resolution. For example, Chinese Patent Application CN108974043A describes an automatic rail detection system that uses a line scan camera and an irradiation or light source that includes light emitting diodes (LEDs) to capture images with clear pixel resolution under varying lighting conditions. Although, line scan cameras may address the issue associated with pixel resolution, modern-day rail track inspection systems still lack techniques that avoid aberration and parallax errors in images captured by the imaging systems. Parallax and aberration errors in the captured images cause existing rail defect identifying systems to identify defects in rail tracks inaccurately, which in turn may lead to train accidents. Further, addition of a source of irradiation, such as LEDs for compensating for the varying lighting conditions, increases power consumption, cost and weight of the imaging systems.
[0006] Accordingly, there remains a need for an improved inspection system that is capable of capturing high contrast images with sufficient resolution and clarity under dynamic varying lighting conditions.

BRIEF DESCRIPTION

[0007] It is an objective of the present disclosure to provide an inspection system. The inspection system includes a line scan camera that includes a single line of sensor pixels and is adapted to capture an initial line scan image of a target object to be inspected, and an image quality management system. The image quality management system that is communicatively coupled to the line scan camera and is adapted to determine a lighting condition prevailing in the surroundings of the target object. Further, the inspection system includes a telecentric liquid lens that is operatively coupled to the line scan camera and the image quality management system. The telecentric liquid lens is adapted to adjust one or more associated lens parameters, the line scan camera is adapted to adjust one or more associated camera parameters, or a combination thereof, based on one or more of the lighting condition determined by the image quality management system and a vertical distance between the line scan camera and the target object for enabling the line scan camera to capture a subsequent set of line scan images of the target object.
[0008] Furthermore, the inspection system includes a defect identification system that is operatively coupled to the line scan camera and is adapted to process the subsequent set of line scan images of the target object captured by the line scan camera to identify if one or more defects exist in the target object. Moreover, the inspection system includes an alerting system that is operatively coupled to the defect identification system and is adapted to transmit an alert message to a ground control station when the defect identification system identifies the defects in the target object. The target object corresponds to one or more rails of a railway track. The defects identified in the target object correspond to one or more of a defective joint gap and a defective switch gap in the rails of the railway track. The inspection system is operatively coupled to an unmanned aerial vehicle. The unmanned aerial vehicle moves over the railway track such that an unmanned vehicle path center axis aligns with a railway track center axis during inspection of the defects in the rails of the railway track. The alerting system transmits the alert message to the ground control station when the defect identification system identifies one or more of the defective joint gap and the defective switch gap in the rails. The alert message corresponds to one of a text alert message, an audio alert message, and a video alert message.
[0009] The alert message specifies types of the defects including one or more of the defective joint gap and the defective switch gap identified in the rails, a width and a geographical location of the defective joint gap, and a width and a geolocation of the defective switch gap. The alerting system transmits the alert message to the ground control station maintained for scheduled maintenance when the width of the defective joint gap or the width of the defective switch gap deviates from a corresponding designated range by less than a specific threshold. The alerting system transmits the alert message to the ground control station for emergency maintenance when the width of the defective joint gap or the width of the defective switch gap deviates from the corresponding designated range by more than the specific threshold. The telecentric liquid lens includes one of an object space telecentric liquid lens, an image space telecentric liquid lens, a double telecentric liquid lens, bi-telecentric liquid lens, and a telecentric motorized lens. The one or more associated lens parameters of the telecentric liquid lens include one or more of a working distance, a lens focal axis, a lens orientation, a lens iris, a focal length, a lens zooming level, and a field of view of the telecentric liquid lens.
[0010] The one or more associated camera parameters of the line scan camera include one or more of a shutter speed, gamma correction, a camera gain, white balance, and a color image format. The inspection system includes one or more of a road defect identifying system, a railway track inspection system, a manufacturing facility, a service station, an aerial photography system, a geographical mapping system, a package delivery system, an agriculture management system, a disaster management system, and a weather forecast system.
[0011] It is an objective of the present disclosure to provide a method for inspecting a target object. The method includes capturing an initial line scan image of the target object using a line scan camera that includes a single line of sensor pixels, and determining a lighting condition prevailing in the surroundings of the target object from the initial line scan image of the target object by an image quality management system that is communicatively coupled to the line scan camera. Further, the method includes adjusting one or more of one or more lens parameters of a telecentric liquid lens and one or more camera parameters of the line scan camera based on one or more of the lighting condition determined by the image quality management system and a vertical distance between the line scan camera and the target object. The method further includes capturing a subsequent set of line scan images of the target object by the line scan camera after adjusting one or more of the one or more lens parameters of the telecentric liquid lens and the one or more camera parameters of the line scan camera. Moreover, the method includes processing the subsequent set of line scan images of the target object captured by the line scan camera by a defect identification system to identify if one or more defects exist in the target object. In addition, the method includes transmitting an alert message to a ground control station by an alerting system when the defect identification system identifies the defects in the target object.
[0012] The target object corresponds to one or more rails of a railway track. The defects identified in the target object corresponds to one or more of a defective joint gap and a defective switch gap in the rails of the railway track. Further, the method includes adjusting one or more lens parameters of the telecentric liquid lens includes adjusting one or more of a focus and a working distance of the telecentric liquid lens based on the vertical distance between the line scan camera and the rails when the initial image of the target object includes one or more contrast issues. Furthermore, the method includes adjusting one or more of the one or more lens parameters of the telecentric liquid lens and one or more camera parameters of the line scan camera based on the determined lighting condition includes identifying that the initial line scan image of the rails captured by the line scan camera is an underexposed image when pixel intensity values of a designated percentage of pixels in the initial line scan image fall in a first pixel intensity range.
[0013] Moreover, the method includes identifying that the initial line scan image of the rails captured by the line scan camera is an overexposed image when pixel intensity values of the designated percentage of pixels in the initial line scan image fall in a second pixel intensity range. In addition, the method includes performing a first corrective action when the initial line scan image of the rails captured by the line scan camera corresponds to the underexposed image. The first corrective action includes one or more of increasing a shutter speed of the line scan camera, increasing a brightness level of the line scan camera, increasing an iris value of the telecentric liquid lens, and increasing a gamma level of the line scan camera. The method further includes performing a second corrective action when the initial line scan image of the rails captured by the line scan camera corresponds to the overexposed image. The second corrective action includes one or more of decreasing a shutter speed of the line scan camera, decreasing a brightness level of the line scan camera, decreasing an iris value of the telecentric liquid lens, and decreasing a gamma level of the line scan camera. Capturing the initial line scan image of the rails using the line scan camera includes adjusting a position of a mechanical aperture stop in the telecentric liquid lens and a position of one or more of an imaginary entrance pupil and an imaginary exit pupil associated with the telecentric liquid lens when the initial line scan image of the rails captured by the line scan camera includes one or more of a parallax error and an aberration error.
[0014] Identifying if one or more defects exist in the rails by the inspection system includes combining the subsequent set of line scan images of the rails captured by the line scan camera into an image of a particular portion of the rails. Further, the method includes converting the image into a greyscale image, performing image segmentation and converting the greyscale image including edges of the rails into a binary image, and detecting a first rail region in the binary image including a joint gap and a second rail region in the binary image including a switch gap using a pattern matching technique. Furthermore, the method includes cropping an image portion including the first rail region from the greyscale image, and generating a first horizontal line and a second horizontal line across the joint gap in the first rail region.
[0015] Moreover, the method includes identifying a first left boundary and a first right boundary of the joint gap lying in the first horizontal line and a first number of pixels that exist between the first left boundary and the second right boundary. In addition, the method includes identifying a second left boundary and a second right boundary of the joint gap lying in the second horizontal line and a second number of pixels that exist between the second left boundary and the second right boundary. Further, the method includes determining an average of the identified first number of pixels and the identified second number of pixels. The determined average corresponds to a width of the joint gap in the first rail region. Furthermore, the method includes identifying that the joint gap in the first rail region as the defective joint gap when the determined width of the joint gap deviates from a first designated range. Identifying if one or more defects exist in the rails by the inspection system includes cropping an image portion including the second rail region from the image to obtain a cropped image.
[0016] Further, the method includes performing a gap segmentation on the cropped image to obtain a gap segmented image from the cropped image and converting the gap segmented image into a color image, and generating a first horizontal line and a second horizontal line across the switch gap in the second rail region. Furthermore, the method includes identifying a first left boundary and a first right boundary of the switch gap lying in the first horizontal line and a first number of pixels that exist between the first left boundary and the second right boundary. Moreover, the method includes identifying a second left boundary and a second right boundary of the switch gap lying in the second horizontal line and a second number of pixels that exist between the second left boundary and the second right boundary. In addition, the method includes determining an average of the identified first number of pixels and the identified second number of pixels.
[0017] The determined average corresponds to a width of the switch gap in the second rail region. The method further includes identifying that the switch gap in the second rail region as the defective switch gap when the determined width of the switch gap deviates from a second designated range. The alert message that is transmitted to the ground control station includes the defects identified in the rails including one or more of the defective joint gap and the defective switch gap, widths of one or more of the defective joint gap and the defective switch gap, and geographical locations of one or more of the defective joint gap and the defective switch gap.

BRIEF DESCRIPTION OF DRAWINGS

[0018] These and other features, aspects, and advantages of the claimed subject matter will become better understood when the following detailed description is read with reference to the accompanying drawings in which like characters represent like parts throughout the drawings, wherein:
[0019] FIG. 1 illustrates a block diagram depicting an exemplary railway track inspection (RTI) system that is operatively coupled to an unmanned aerial vehicle (UAV) for accurately identifying defects in one or more rails of a railway track, in accordance with aspects of the present disclosure;
[0020] FIG. 2A illustrates an exemplary illustration depicting a switch region including a switch gap in the rails;
[0021] FIG. 2B illustrates another illustration depicting a joint region including a joint gap in the rails;
[0022] FIG. 3 illustrates a flow diagram depicting an exemplary method for capturing high contrast and optimally exposed images of the rails using an image quality system in the RTI system of FIG. 1, in accordance with aspects of the present disclosure;
[0023] FIG. 4A illustrates an exemplary illustration of a camera that is operatively coupled to the UAV of FIG. 1 for capturing images of the rails;
[0024] FIG. 4B illustrates an exemplary illustration depicting an image including parallax error that is captured using a conventional area scan camera;
[0025] FIG. 4C illustrates an exemplary illustration depicting an image that is free from parallax error and is captured using a line scan camera in the RTI system of FIG. 1, in accordance with aspects of the present disclosure;
[0026] FIG. 4D illustrates an exemplary illustration depicting an image including aberration error that is captured using a conventional area scan camera;
[0027] FIG. 4E illustrates an exemplary illustration depicting an image that is free from aberration error and is captured using the line scan camera in the RTI system of FIG. 1, in accordance with aspects of the present disclosure;
[0028] FIG. 5 illustrates an exemplary illustration depicting the UAV that is oriented in a particular position with respect to the rails during inspection of defects in the rails, in accordance with aspects of the present disclosure;
[0029] FIG. 6A illustrates an exemplary illustration depicting the conventional area scan camera that is operatively coupled to an UAV for capturing images of the rails area-wise;
[0030] FIG. 6B illustrates an exemplary illustration of the line scan camera that is operatively coupled to the UAV of FIG. 1 for capturing images of the rails line-by-line, in accordance with aspects of the present disclosure;
[0031] FIG. 7A illustrates a graphical representation view depicting a histogram associated with an underexposed image of the rails;
[0032] FIG. 7B illustrates a graphical representation view depicting a histogram associated with an overexposed image of the rails;
[0033] FIG. 7C illustrates a graphical representation view depicting a histogram associated with a high contrast and an optically exposed image of the rails;
[0034] FIG. 8 illustrates an exemplary illustration depicting an image of a portion of the rails including the switch gap;
[0035] FIGS. 9A-B illustrate a flow diagram depicting an exemplary method used by a rail defect identification system in the RTI system of FIG. 1 to identify defective switch and joint gaps in the rails, in accordance with aspects of the present disclosure;
[0036] FIG. 10A illustrates an exemplary illustration depicting an image of a particular portion of the rails generated by the rail defect identification system of FIG. 1 by combining line-by-line high contrast and optimally exposed images of the rails captured by the line scan camera, in accordance with aspects of the present disclosure;
[0037] FIG. 10B illustrates an exemplary illustration depicting a greyscale image that is derived from the image depicted in FIG. 10A, in accordance with aspects of the present disclosure;
[0038] FIG. 10C illustrates an exemplary illustration depicting a binary image that is generated from the greyscale image depicted in FIG. 10B, in accordance with aspects of the present disclosure;
[0039] FIG. 10D illustrates an exemplary illustration depicting a first rail region including a joint gap and a second rail region including a switch gap in the rails that are detected from the binary image of FIG. 10C, in accordance with aspects of the present disclosure;
[0040] FIG. 10E illustrates an exemplary illustration depicting a greyscale image that is generated from the binary image of FIG. 10C, in accordance with aspects of the present disclosure;
[0041] FIG. 11A illustrates an exemplary illustration depicting an image of the second rail region including the switch gap that is cropped from the image depicted in FIG. 10A, in accordance with aspects of the present disclosure;
[0042] FIG. 11B illustrates an exemplary illustration depicting a gap segmented image that is generated from the cropped image of FIG. 11A using the rail defect identification system, in accordance with aspects of the present disclosure; and
[0043] FIG. 11C illustrates an exemplary illustration depicting a color image that is generated from the gap segmented image of FIG. 11B using the rail defect identification system, in accordance with aspects of the present disclosure.

DETAILED DESCRIPTION

[0044] The following description presents an exemplary inspection system that captures high contrast and underexposure and/or overexposure defects free images of rails of a railway track in dynamically changing ambient lighting conditions, and identifies defects in the rails from the captured images accurately. Particularly, embodiments described herein disclose the inspection system that accurately identifies defective switch gaps and joint gaps in the rails of the railway track in different ambient lighting conditions. Further, the inspection system transmits alert messages including geolocations of the identified defective switch and/or joint gaps to a ground control station for enabling maintenance personnel to reach the geographical locations and carry out repair and/or maintenance activities in a timely manner.
[0045] As noted previously, railway tracks require periodic inspections to identify associated defects and to prevent derailing of trains, which leads to accidents. Certain conventional railway track inspection systems employ an unmanned aerial vehicle such as a drone for identifying defects in the railway track. Such conventional systems generally include an area scan camera and a fixed focus lens that are mounted to the drone. The drone, carrying the area scan camera and the fixed focus lens, moves over the railway track such that the area scan camera continuously captures images of the railway track. The conventional systems then process the images captured by the area scan camera and identify defects in the rails from the captured images.
[0046] However, the area scan camera used in the conventional systems for capturing images of the rails generally provides low pixel resolution images as outputs. Consequently, the conventional systems may not be able to accurately identify defects such as defective switch and joint gaps in the rails from such low-resolution images. Further, the fixed focus lens used in the conventional systems causes the area scan camera to capture images of the rails that include parallax and aberration errors. In addition, the fixed focus lens is not capable of adjusting an associated focus based on a height between the drone carrying the area scan camera and the rails, thus causing the area scan camera to capture images that include contrast issues. As a result, the conventional systems identifying the defects in the rails using such images including parallax, aberration and contrast errors fail to accurately identify the defects in the rails.
[0047] Furthermore, parameters of the area scan camera and the fixed focus lens used in the conventional systems are not typically variable, and thus, cannot be varied according to dynamically changing ambient lighting conditions. This inability of the conventional systems to vary the parameters of the area scan camera and/or fixed focus lens causes the images captured by the area scan camera to be underexposed or overexposed images according to the prevailing ambient lighting condition.
[0048] For example, the images captured by the area scan camera may be underexposed images when those images were captured during evening-time, nighttime, and any other dark lighting condition. In another example, the images captured by the area scan camera may be overexposed images when those images were captured in daytime or bright lighting conditions. Such underexposed or overexposed images either lead to missing identification of actual defects in the rails that may cause train accidents or leads to sharing of false alert messages to the ground control station even though no actual defects exist in the rails.
[0049] In order to address the aforementioned issues, the present disclosure provides an efficient inspection system that employs a line scan camera and telecentric liquid lens instead of the area scan camera and the fixed focus lens used in the conventional systems. The line scan camera employed by the inspection system captures and outputs high-resolution images of the rails of the railway track. The inspection system is configured to accurately determine widths of switch and/or joint gaps in the rails from the high-resolution images captured by the line scan camera.
[0050] Further, the telecentric liquid lens employed by the inspection system dynamically adjusts associated lens parameters such that the images captured by the line scan camera are free from parallax and aberration errors. The inspection system, thus identifying defects in the rails of the railway track from the parallax and aberration errors free images, is able to accurately identify the defects in the rails. Further, the line scan camera and the telecentric liquid lens employed by the inspection system allow the inspection system to adjust associated camera and/or lens parameters, for example, according to dynamically changing ambient lighting conditions and/or a ground-to-aerial distance between the line scan camera and the railway track.
[0051] Accordingly, the inspection system described in the present disclosure is able to consistently capture high contrast and optimally exposed images irrespective of lighting conditions prevailing in the outdoor environment. Further, the inspection system processes such high contrast and optimally exposed images, and accurately identifies defects in the rails in dynamically changing ambient lighting conditions. Upon accurately identifying defects in the rails, the inspection system transmits alert messages to the ground control station for enabling the maintenance personnel to carryout timely track repair and/or maintenance activities, and thereby prevents derailing of trains and related accidents, and further ensures safety of the passengers.
[0052] It may be noted that different embodiments of the present inspection system may be used in different application areas. For example, the inspection system may be used in a road defect identifying system to identify defects such as cracks, potholes, and sinkholes on roads, and to transmit an alert message including geolocations of such defective road portions to a concerned authority. In another example, the inspection system may be used in a manufacturing facility or a service station in identifying defects in vehicle tires such as shrinking and enlargement of the vehicle tires at associated corner and center portions, respectively. Also, the inspection system identifies and validates characters such as batch or model names of vehicle tires and a manufacturing date of the vehicle tires mentioned on the vehicle tires based on 3-dimensional (3D) information of the vehicle tires, for example, determined using a point cloud technique. In yet another example, the inspection system may be used in an aerial photography system to capture high resolution aerial photos of various geographical regions. The inspection system may also be used in a geographical mapping system to acquire high resolution images of difficult-to-reach locations such as coastlines, mountaintops, and islands, and to create 3-D maps of such locations.
[0053] Furthermore, the inspection system may be used in a package delivery system to deliver a package via air using an unmanned aerial vehicle and to capture a photo of a person who collected the package. The inspection system may be used in an agriculture management system to monitor the health of crops and to monitor application of fertilizers and/or insecticides on the crops by an unmanned aerial vehicle in real-time. The inspection system may be used in a disaster management system to locate one or more victims of a disaster such as flood, earthquake, a fire accident, and a landslide incident.
[0054] The inspection system may also be used in a weather forecast system to capture images of hurricanes and tornadoes, and to acquire new insights related to behavior and trajectories of such hurricanes and tornadoes. For clarity, an embodiment of the present inspection system is described herein in greater detail with reference to a railway track inspection system configured to identify defects in rails of a railway track and to initiate automated actions, as and when needed.
[0055] FIG. 1 illustrates a block diagram depicting an exemplary railway track inspection (RTI) system (100) that is operatively coupled to an unmanned aerial vehicle (UAV) (102) for accurately identifying defects in a target object (104) such as in one or more rails (104) of a railway track (106). In certain embodiments, the RTI system (100) accurately identifies defects in the rails (104) irrespective of dynamically changing lighting conditions in the surroundings of the rails (104). The lighting conditions in the surroundings of the rails (104) change depending upon various factors such as depending upon a particular time of the day such as daytime, noontime, evening-time, or nighttime, a number of artificial lights lit in the surroundings of the rails (104), shadows casted over portions of the rails (104), and/or if the rails (104) correspond to underground rails or outdoor rails. Examples of the defects in the rails (104) that may be identified using the RTI system (100) include defective switch gaps, defective joint gaps, rail cracks, rail squats, and rail flaking in the rails (104). Though the RTI system (100) is capable of identifying different types of defects in the rails (104), the present embodiment describes identification of certain specific types of defects such as the defective switch gaps and the defective joint gaps in the rails (104) by the RTI system (100) in greater detail with reference to FIGS. 2A-11C.
[0056] For example, FIG. 2A illustrates an exemplary illustration depicting a switch region (202) in the rails (104). The switch region (202) in the rails (104) generally includes a first tongue track (204A) and a second tongue track (204B). In this example, a gap that exists between the second tongue track (204B) and one of the main rails (104) corresponds to a switch gap (206). The switch gap (206) in the rails (104) typically is used to perform an important functionality of diverting a train to another intended track and direction.
[0057] Generally, a width of the switch gap (206) needs to be within a predetermined range in order to properly divert a train to another intended track. However, in real-world scenarios, the width of the switch gap (206) is not constant and varies over a period of time due to continuous mechanical friction between the tongue tacks (204A-B) and wheels of the train. When the width of the switch gap (206) in the switch region (202) deviates from the predetermined range, the train cannot be properly diverted to another intended track, often leading to derailment. Accordingly, the RTI system (100) can be used to identify such defective switch gaps in the rails (104) whose associated widths deviate from the predetermined range.
[0058] Similarly, FIG. 2B illustrates another illustration depicting a joint region (208) in the rails (104). The joint region (208) typically includes a joint gap (210) that exists between two subsequent pieces of the rail (104). Typically, a width of the joint gap (210) needs to be always within a designated range in order to prevent train accidents. However, in real-world scenarios, the width of the joint gap (210) often deviates from the designated range due to local deformation and/or bulging of the rail (104) that occurs due to thermal stress, expansion, and/or shrinking of the rail (104). Accordingly, the RTI system (100) can be used to identify such defective joint gaps in the rails (104) whose associated widths deviate from the predetermined range.
[0059] Referring back to FIG. 1, in one embodiment, the RTI system (100) includes a set of components including a telecentric liquid lens (108), a line scan camera (110), an image quality management system (112), a defect identification system (114) such as a rail defect identification system (114), a location sensor (116), a power unit (118), and an alerting system (126) in order to identify such defective switch and joint gaps in the rails (104). Particularly, in contrast to conventional inspection systems that use a fixed focus lens with fixed parameters, the RTI system (100) uses the telecentric liquid lens (108) whose associated lens parameters are adjustable. Examples of the telecentric liquid lens (108) employed by the RTI system (100) include an object space telecentric liquid lens, an image space telecentric liquid lens, double or bi-telecentric liquid lens, and a telecentric motorized lens.
[0060] Generally, the telecentric liquid lens (108) is an imaging lens that optically corrects perspective errors that occur due to parallax. Further, the telecentric liquid lens (108) quickly adjusts a lens focus to accommodate and properly focus onto regions of interest of the rails (104) located at different working distances. Additionally, the telecentric liquid lens (108) provides a constant magnification of the rails (104) over a wide range of working distances for virtually eliminating viewing angle errors. To that end, the telecentric liquid lens (108) includes mechanically or electrically controlled cells including optical-grade liquid.
[0061] In one embodiment, the image quality management system (112) applies current or voltage signals to the cells of the telecentric liquid lens (108) to suitably change shapes of one or more of the cells. The changes in the shapes of the cells adjust one or more lens parameters associated with the telecentric liquid lens (108). Examples of the lens parameters that may be adjusted by the image quality management system (112) by application of current or voltage signals include a field of view, a focal length, a working distance, a lens focal axis, a lens orientation, a lens iris value, and a lens zooming level of the telecentric liquid lens (108).
[0062] In certain embodiments, the image quality management system (112) adjusts the lens parameters of the telecentric liquid lens (108) such that the line scan camera (110) captures parallax and aberration free images of the rails (104), as described in detail with reference to FIG. 3. Capturing of such parallax and aberration free images of the rails (104) by the line scan camera (110) enables the rail defect identification system (114) to accurately determine defective switch and/or joint gaps in the rails (104).
[0063] Subsequent to adjusting the one or more lens parameters of the telecentric liquid lens (108), the line scan camera (110) captures an image of a selected region of interest in the rails (104), and provides the captured image as an input to the image quality management system (112). The image quality management system (112) then determines a lighting condition prevailing in the outdoor environment by performing a histogram analysis of the image. For instance, the image quality management system (112) identifies that the lighting condition prevailing in the outdoor environment corresponds to a bright lighting condition such as a daytime and/or a noontime lighting condition when a majority, for example, 80 percentage of pixels in the image includes associated pixel intensities in the range of 200-256. Generally, the image captured under such bright lighting condition will be an overexposed image having oversaturation and contrast issues. When the rail defect identification system (114) identifies widths of switch and joint gaps in the rails (104) from such an overexposed image, the rail defect identification system (114) may not identify defective switch and joint gaps accurately.
[0064] In another example, the image quality management system (112) identifies that the lighting condition prevailing in the outdoor environment corresponds to a dark lighting condition such as an evening-time and/or a nighttime lighting condition when the majority of pixels in the image having associated pixel intensities in the range of 0-50. Generally, the image captured under such dark lighting condition will be an underexposed image having undersaturation and poor contrast issues. When the rail defect identification system (114) identifies widths of switch and joint gaps in the rails (104) from such an underexposed image, the rail defect identification system (114) may not identify defective switch and joint gaps accurately.
[0065] In order to address the aforementioned issues and to capture consistent, uniform intensity, and high contrast images of the rails (104), the image quality management system (112) adjusts one or more lens parameters associated with the telecentric liquid lens (108) and/or one or more camera parameters associated with the line scan camera (110) based on the determined lighting condition prevailing in the environment. Examples of the one or more camera parameters include a shutter speed, gamma correction, a camera gain, white balance, and a color format of the images to be captured by the line scan camera (110).
[0066] According to aspects of the present disclosure, the one or more lens parameters and/or the one or more camera parameters are continually adjusted according to the prevailing lighting condition as the UAV (102) continues to move over the railway track (106) such that the line scan camera (110) coupled to the UAV (102) continuously captures a plurality of images of the rails (104). The line scan camera (110), thus, is able to capture high contrast and optimally exposed images that are suitable for identifying defective switch and joint gaps in the rails (104) accurately. In certain embodiments, the rail defect identification system (114) combines the plurality of images captured by the line scan camera (110) into a combined image. Further, rail defect identification system (114) processes the combined image and identifies defective switch and/or joint gaps in the rails (104) using an exemplary methodology described in greater detail with reference to FIGS. 9A-B.
[0067] In certain embodiments, the location sensor (116) in the RTI system (100) identifies geographical locations of switch and/or joint gaps that are identified to be defective by the rail defect identification system (114). Upon identifying geographical locations of such defective switch and/or joint gaps, the alerting system (126) transmits an alert message to a ground control station (120) via a communications link (122). In one embodiment, the ground control station (120) corresponds to a remote computer or a server that is accessible by a railway track repair and maintenance division. Examples of the communications link (122) via which the alert message is shared with the ground control station (120) include a Wi-Fi network, an Ethernet, and a cellular data network.
[0068] In certain embodiments, the alert message transmitted to the ground control station (120) includes geographical locations of the defective switch and/or joint gaps in the rails (104) for enabling the railway track repair and maintenance division to send technicians to those geographical locations and to carry out repair and/or maintenance activities. In one embodiment, the location sensor (116) and the rail defect identification system (114) that identifies defective switch and/or joint gaps in the rails (104) resides in the RTI system (100), as depicted in FIG. 1.
[0069] However, it is to be understood that the location sensor (116) and the rail defect identification system (114), and the alerting system (126) may alternatively reside in the UAV (102) instead of residing in the RTI system (100). In this embodiment, the rail defect identification system (114) residing in the UAV (102) receives the plurality of images captured by the line scan camera (110) in the RTI system (100), and combines those images into the combined image. Further, the rail defect identification system (114) in the UAV (102) identifies one or more defective switch and/or joint gaps from the combined image, and receives information including geographical locations where images of such defective switch and/or joint gaps were captured from the location sensor (116) in the UAV (102). Subsequently, the alerting system (126) in the UAV (102) transmits the alert message that specifies the nature of defects identified in the rails (104) such as defective switch and/or joint gaps identified in the rails (104) and corresponding geographical locations to the ground control station (120) via an associated communication unit (124). Examples of the communication unit (124) include a Wi-Fi network, an Ethernet, and a cellular data network.
[0070] In another embodiment, the location sensor (116) resides either in the RTI system (100) or in the UAV (102), whereas the rail defect identification system (114) and the alerting system (126) reside in the ground control station (120). In this embodiment, the image quality management system (112) transmits the plurality of images captured by the line scan camera (110) along with geographical locations where those images were captured to the rail defect identification system (114) in the ground control station (120). The rail defect identification system (114) then combines the plurality of images into the combined image. Further, the rail defect identification system (114) identifies defective switch and/or joint gaps in the rails (104) from the combined image along with their respective geographical locations for enabling the railway track repair and maintenance division to carry out repair and/or maintenance activities.
[0071] In one embodiment, the rail defect identification system (114) and the image quality management system (112), for example, may include one or more general-purpose processors and specialized processors to capture high contrast and optimally exposed images and to accurately identify defective switch and/or joint gaps in the rails (104) irrespective of lighting conditions prevailing in the outdoor environment. In certain embodiments, the rail defect identification system (114) and the image quality management system (112) may also include one or more graphical processing units, microprocessors, programming logic arrays, field programming gate arrays, integrated circuits, system on chips, and/or other suitable computing devices. Accordingly, certain operations of the rail defect identification system (114) and the image quality management system (112) may be implemented by suitable code on a processor-based system, such as a general-purpose or a special-purpose computer.
[0072] In certain embodiments, the RTI system (100) that is operatively coupled to the UAV 102) includes the power unit (118) that supplies power to the telecentric liquid lens (108), the line scan camera (110), and the location sensor (116) for associated operations. An example of the power unit (118) that supplies power to components in the RTI system (100) includes a battery and a solar cell. An exemplary methodology used for capturing high contrast and optimally exposed images of the rails (104) irrespective of lighting conditions prevailing in the outdoor environment is described in detail with reference to FIG. 3.
[0073] FIG. 3 illustrates a flow diagram depicting an exemplary method (300) for capturing high contrast and optimally exposed images of the rails (104) using the image quality management system (112) in the RTI system (100). In one embodiment, the image quality management system (112) receives desired contrast and exposure values corresponding to the captured images from a user. Alternatively, the image quality management system (112) determines desired image contrast and exposure levels to generate high contrast and optimally exposed images based on previously processed images. The order in which the exemplary method is described is not intended to be construed as a limitation, and any number of the described blocks may be combined in any order to implement the exemplary method disclosed herein, or an equivalent alternative method. Additionally, certain blocks may be deleted from the exemplary method or augmented by additional blocks with added functionality without departing from the claimed scope of the subject matter described herein.
[0074] At step (302), the line scan camera (110) captures an initial image of a selected region of interest of the rails (104) of the railway track (106). In one embodiment, the image quality management system (112) identifies if the initial image captured by the line scan camera (110) includes parallax and aberration errors. In certain embodiments, the parallax error occurs in the initial image when a horizontal axis (402) of the line scan camera (110) is not positioned exactly parallel to a horizontal axis (404) of the rails (104) during capturing of the initial image by the line scan camera (110), as depicted in FIG. 4A. In other words, the UAV (102) carrying the line scan camera (110) is expected to align and move exactly parallel to the rails (104) such that the horizontal axis (402) of the line scan camera (110) is positioned exactly parallel to the horizontal axis (404) of the rails (104) for enabling the line scan camera (110) to capture parallax free images of the rails (104). However, in real-world scenarios, the UAV (102) does not always move exactly parallel to the rails (104) and often tilts when there is a strong wind. The tilting of the UAV (102) causes tilting of the line scan camera (110) as the line scan camera (110) is attached to the UAV (102). The tilting of the line scan camera (110) causes the horizontal axis (402) of the line scan camera (110) to not be parallel with the horizontal axis (404) of the rails (104), thereby introducing parallax errors in the captured images.
[0075] For example, FIG. 4B depicts an exemplary image (406) of a set of objects that is captured when the horizontal axes (402 and 404) are not parallel with respect to each other. As evident from FIG. 4B, the image (406) includes a parallax error in which magnification levels of the objects represented as dark circles in the image (406) are not uniform. In one embodiment, the objects in FIGS. 4B-E represent specific sections of the rails (104) of the railway track (106). For the sake of simplicity, the specific rail sections are depicted as dark circles in FIGS. 4B-E to depict and describe that the images captured using an area scan camera and a fixed focus lens generally include parallax and aberration errors. However, in real-world scenarios, these specific rail sections will be appearing different and do not appear as dark circles when images of the specific rail sections are captured using a camera such as the line scan camera (110). When the rail defect identification system (114) identifies defects in the rails (104) from such an image (406) affected by the parallax error, the rail defect identification system (114) may not identify defective switch and/or joint gaps accurately.
[0076] In certain embodiments, the aberration error occurs in captured images when a lens used for capturing those images includes an angular field of view such that a chief ray of the lens is not parallel to an optical axis of the lens. In one embodiment, the chief ray refers to a ray from an off-axis point in an object passing through a center of a mechanical aperture stop in the telecentric liquid lens (108). For example, certain conventional railway track inspection systems use a fixed focus lens that include an angular field of view, and hence, the images captured using such fixed focus lens may include aberration errors. For example, FIG. 4D depicts one such exemplary image (410) that is captured using the conventional fixed focus lens and includes an aberration error.
[0077] Specifically, the image (410) includes the aberration error in which the magnification levels of objects in a center region (412) of the image (410) are different from the magnification levels of objects in a corner region (414) of the image (410). Further, the orientations of the objects in the corner region (414) are misplaced. Additionally, the distance between one object to another object in the center region (412) is comparatively greater than the distance between one object to another object in the corner region (414). When the rail defect identification system (114) identifies defects in the rails (104) from such an image (410) affected by the aberration error, the rail defect identification system (114) may not identify defective switch and/or joint gaps accurately.
[0078] At step (304), the image quality management system (112) dynamically adjusts the one or more lens parameters associated with the telecentric liquid lens (108) when the initial image captured by the line scan camera (110) includes parallax and/or aberration errors. Dynamically adjusting the one or more lens parameters of the telecentric liquid lens (108) enables the line scan camera (110) to consistently capture a subsequent set of images of the rails (104) that are free from parallax and aberration errors irrespective of dynamically changing ambient lighting conditions.
[0079] For example, the image quality management system (112) adjusts the lens parameters such as a position of a mechanical aperture stop in the telecentric liquid lens (108) and/or a position of one or more of an imaginary entrance pupil and an imaginary exit pupil associated with the telecentric liquid lens (108) such that the line scan camera (110) captures the subsequent set of images of the rails (104) that are free from parallax errors. For instance, FIG. 4C depicts another image (408) that is captured by the line scan camera (110) after adjusting one or more of the lens parameters of the telecentric liquid lens (108). Unlike the magnification levels of objects in the image (406) depicted in FIG. 4B, the magnification levels of objects in the image (408) are all uniform and the image (408) is free from parallax error. The parallax free images, thus captured using the telecentric liquid lens (108) and the line scan camera (110), enable the rail defect identification system (114) to identify defective switch and/or joint gaps in the rails (104) accurately in real-world scenarios.
[0080] Similar to enabling the line scan camera (110) to capture the subsequent set of images of the rails (104) that are free from parallax error, the image quality management system (112) also enables the line scan camera (110) to capture the subsequent set of images of the rails (104) that are free from aberration errors. To that end, the image quality management system (112) employs the telecentric liquid lens (108), which is a zero angular field of view or a non-angular field of view type lens. Specifically, the image quality management system (112) adjusts the lens parameters such as a position of a mechanical aperture stop in the telecentric liquid lens (108) and/or a position of one or more of an imaginary entrance pupil and an imaginary exit pupil associated with the telecentric liquid lens (108). Adjusting one or more of these lens parameters disposes the chief ray of the telecentric liquid lens (108) to be parallel with the optical axis of the telecentric liquid lens (108) and enables the line scan camera (110) to capture the subsequent set of images of the rails (104) that are free from aberration errors.
[0081] For example, FIG. 4E depicts an image (416) that is free from aberration error and is captured by the line scan camera (110) after adjusting one or more of the lens parameters of the telecentric liquid lens (108). As it may be noted from FIG. 4E, the magnification levels of objects in all regions of the image (416) are the same, orientations of the objects in a corner region (418) are not misplaced, and the distance between one object to another object in any region of the image (416) is same. When the rail defect identification system (114) captures aberration free images similar to the image (416), the rail defect identification system (114) is able to identify defective switch and/or joint gaps in the rails (104) accurately.
[0082] In certain embodiments, the UAV (102) carrying the line scan camera (110) may tilt and orient upside down such that the line scan camera (110) may not be able to capture images of the rails (104) when there is a strong wind. In order to prevent such upside down position of the UAV (102) during strong wind conditions, the UAV (102) includes an integrated rotor speed control system (not shown in FIGS. 1-11C). During strong wind conditions, the rotor speed control system controls speeds such as increase and/or decrease speeds of one or more rotors of the UAV (102) based on the wind direction and force before the UAV (102) flips upside down completely and orients the UAV (102) back to an original non-titled position for enabling the line scan camera (110) to continue to capture images of the rails (104).
[0083] In certain other embodiments, the UAV (102) may be pushed away and displaced from a current position when there is a strong wind, which may cause the line scan camera (110) to not be able to capture images of both the rails (104). For example, FIG. 5 illustrates an exemplary illustration depicting the UAV (102) that is oriented in a particular position (502) above the rails (104) such that a center axis of a unmanned vehicle path (unmanned vehicle path center axis) (504) is aligned with a center axis of a railway track (railway track center axis) (506). The UAV (102), thus oriented in the particular position (502), captures images (508) of both the rails (104) of the railway track (106). However, the UAV (102) moves away from the particular position (502) to another position, for example, a displaced position (510) when there is a strong wind. In such a scenario, the line scan camera (110) may be able to capture images (512) of only one of the rails (104).
[0084] Further, in such a scenario, the unmanned vehicle path center axis (504) misaligns with the railway track center axis (506) horizontally by a certain distance (514) that corresponds to a horizontal distance moved by the UAV (102) from the particular position (502) to the displaced position (510). In certain embodiments, the UAV (102) includes one or more inertial measurement unit (IMU) sensors such as pitch, roll, and yaw sensors (not shown in FIGS. 1-11C) that measure the horizontal distance (514) moved by the UAV (102) from the particular position (502) to the displaced position (510). Subsequently, the rotor speed control system in the UAV (102) controls speeds of one or more rotors of the UAV (102) to move the UAV (102) by the measured distance and to realign the unmanned vehicle path center axis (504) with the railway track center axis (506) such that the line scan camera (110) captures images of both the rails (104) of the railway track (106).
[0085] Conventional inspection systems use an area scan camera that differs from the line scan camera (110) in functioning and performance when capturing images of the rails (104). The area scan camera used in the conventional inspection systems includes certain limitations that render it sub-optimal for accurate identification of railway track defects in dynamically varying lighting conditions. For example, a horizontal resolution and a vertical resolution of a typical area scan camera is generally 8192 pixels and 5460 pixels, respectively. However, the horizontal resolution and the vertical resolution of the line scan camera (110) corresponds to 16532 pixels and an unlimited vertical resolution, respectively. Hence, the line scan camera (110) outputs high pixel resolution images when compared to resolutions of images output by the area scan camera. The high pixel resolution images output by the line scan camera (110) helps the rail defect identification system (114) to determine widths of switch and/or joint gaps in the rails (104) accurately even when such widths are only few millimeters in real-world scenarios.
[0086] Further, a scanning speed of the typical area scan camera is 500 megahertz (MHz). However, the scanning speed of the line scan camera (110) is 1400 MHz, which is greater than the scanning speed of the area scan camera. Hence, the RTI system (100) completes inspection of the rails (104) quicker when using the line scan camera (110) instead of the area scan camera. Additionally, the area scan camera includes a single rectangular shaped imaging sensor that captures images of the rails (104) area-wise.
[0087] For example, FIG. 6A illustrates an exemplary illustration depicting an area scan camera (602) that is operatively coupled to a UAV (604) for capturing images of the rails (104) area-wise. When the UAV (604) carrying the area scan camera (602) moves over the rails (104), the area scan camera (602) captures a first image of a first rail area (606A). As the UAV (604) continues to move over the rails (104), the area scan camera (602) captures images of subsequent rail areas such as a second image of a second rail area (606B) and a third image of a third rail area (606C). The area scan camera (602) then stitches the captured first, second, and third images into a stitched image (608). Subsequently, the area scan camera (602) transmits the stitched image (608) to the ground control station (120) when the rail defect identification system (114) resides in the ground control station (120) for defect analysis. It may be noted that the area scan camera (602) transmits stitched images similar to the stitched image (608) to the ground control station (120) at a much lower data transfer rate. This is because, the area scan camera (602) captures images of the rails (104) area-wise slowly at the scanning speed of 500 MHz, stitches all the captured images into stitched images, and then transmits the stitched images to the ground control station (120), where all these operations require significant amount of time. The low data transfer rate of the area scan camera (602) increases an overall time taken to inspect and identify defects in the rails (104).
[0088] In contrast to the area scan camera (602), the line scan camera (110) includes comparatively superior data transfer rate. To that end, the line scan camera (110) includes a single row of sensors that capture images of the rails (104) line-by-line in one-pixel slices captured in rapid succession. For example, FIG. 6B illustrates an exemplary illustration of the line scan camera (110) that is operatively coupled to the UAV (102) for capturing images of the rails (104) line-by-line. When the UAV (102) carrying the line scan camera (110) moves over the rails (104), the line scan camera (110) captures a plurality of images (610A-F) of the rails (104) line-by-line as depicted in FIG. 6B. It may be noted that the line scan camera (110) captures the images (610A-F) at an increased scanning speed of 1400 MHz. Further, the line scan camera (110) is generally capable of capturing a next image while transferring a previous image captured by the line scan camera (110) to the ground control station (120). For example, the line scan camera (110) captures the next image (610B) while transferring the previous image (610A) captured by the line scan camera (110) to the ground control station (120).
[0089] This ability of capturing images of the rails (104) at an increased scanning speed and transferring previously captured images while simultaneously acquiring new images enables the line scan camera (110) to provide a superior data transfer rate. Further, each of the images (610A-F) captured by the line scan camera (110) includes unique information such that there is no overlap or redundant information captured in the images (610A-F). In contrast, the images of the rails (104) captured by the area scan camera (602) generally include redundant information. For example, when the area scan camera (602) captures the image of the second rail area (606B), an image of a particular portion (612) in the first rail area (606A) may also be captured in the image of the second rail area (606B). This redundant information in the images captured by the area scan camera (602) makes stitching of those images difficult. Further, transmission of the images including redundant information captured by the area scan camera (602) to the ground control station (120) requires significantly more network bandwidth. In contrast, transmission of the images (610A-F) captured by the line scan camera (110) to the ground control station (120) requires a comparatively lesser network bandwidth as the images (610A-F) captured by the line scan camera (110) are free from redundant information. Further, the area scan camera (602) does not generally acquire images that specifically include only images of the rails (104). The images captured by the area scan camera (602) usually include unnecessary information such as images of objects that exist on left and right sides of the rails (104). The unnecessary information included in the images increases bandwidth consumption and affects the data transfer rate. In contrast, the line scan camera (110) is capable of capturing images of the rails (104) without any unnecessary information included in those images. Hence, the RTI system (100) employing the line scan camera (110) is capable of transmitting images of the rails (104) to the ground control station (120) at much faster rate with lesser consumption of network bandwidth.
[0090] At step (306), the image quality management system (112) determines an ambient lighting condition prevailing in the surroundings of the rails (104) by performing a histogram analysis on the initial image captured using the line scan camera (110). For example, the image quality management system (112) performs the histogram analysis on the captured image and generates a histogram whose associated X-axis represents pixel intensities that range between 0 to 256 and associated Y-axis represents a number or percentage of pixels in the captured image, as depicted in FIGS. 7A-C. In certain embodiments, the image quality management system (112) identifies that the ambient lighting condition prevailing in the surroundings of the rails (104) corresponds to a dark lighting condition such as an evening-time or a nighttime lighting condition when a majority of pixels in the captured image having associated pixel intensities in the range of 0-50, as illustrated in the histogram (700) depicted in FIG. 7A. As noted previously, the image captured under the dark lighting condition will be an underexposed image having undersaturation and poor contrast issues. Hence, the image cannot be used for identifying defects in the rails (104) accurately.
[0091] Alternatively, the image quality management system (112) identifies that the ambient lighting condition prevailing in the surroundings of the rails (104) corresponds to a bright lighting condition such as a daytime lighting condition when a majority of pixels in the captured image having associated pixel intensities in the range of 200-256, as illustrated in the histogram (702) depicted in FIG. 7B. As noted previously, the image captured under the bright lighting condition may be an overexposed image having oversaturation and contrast issues. Hence, such an image also cannot be used for identifying defects in the rails (104) accurately.
[0092] In order to address the aforementioned issues, at step (308), the image quality management system (112) adjusts the one or more lens parameters associated with the telecentric liquid lens (108) and/or the one or more camera parameters associated with the line scan camera (110) based on the determined ambient lighting condition. For example, the image quality management system (112) performs a first corrective action when the determined ambient lighting condition corresponds to the dark lighting condition and when the image captured by the line scan camera (110) corresponds to the underexposed image. Examples of the first corrective action include one or more of increasing a shutter speed of the line scan camera (110), increasing a brightness level of the line scan camera (110), increasing an iris value of the telecentric liquid lens (108), and increasing a gamma level of the line scan camera (110). By performing the first corrective action, the image quality management system (112) enables the line scan camera (110) to capture the subsequent set of images that are high contrast and optimally exposed images and are free from undersaturation and poor contrast issues.
[0093] For example, FIG. 7C illustrates a graphical representation view depicting a histogram (704) associated with an exemplary image captured by the line scan camera (110) after performing the first corrective action by the image quality management system (112). It may be noted that the pixels in the histogram (704) are evenly distributed over all ranges of pixel intensities. Further, it may be noted that the histogram (704) is different from the histogram (700) corresponding to the underexposed image where a majority of pixels are distributed towards the left side of the histogram (700). In addition, the histogram (704) is also different from the histogram (702) corresponding to the overexposed image where a majority of pixels are distributed towards the right side of the histogram (702). The histogram (704) associated with the image including the pixels that are evenly distributed over all ranges of pixel intensities represents that the image captured by the line scan camera (110) after performing the first corrective actions is a high contrast and an optimally exposed image without any saturation and contrast issues. Thus, the image quality management system (112) enables the line scan camera (110) to capture the subsequent set of images that are high contrast and optimally exposed images even in dark lighting conditions by performing the first corrective action.
[0094] Alternatively, the image quality management system (112) performs a second corrective action different from the first corrective action when the determined ambient lighting condition corresponds to the bright lighting condition and when the initial image captured by the line scan camera (110) corresponds to the overexposed image. Examples of the second corrective action include one or more of decreasing a shutter speed of the line scan camera (110), decreasing a brightness level of the line scan camera (110), decreasing an iris value of the telecentric liquid lens (108), and decreasing a gamma level of the line scan camera (110). By performing the second corrective action, the image quality management system (112) enables the line scan camera (110) to capture the subsequent set of images that are high contrast and optimally exposed images and are free from oversaturation and contrast issues even in extremely bright lighting conditions.
[0095] In certain embodiments, the image quality management system (112) also adjusts the one or more lens parameters associated with the telecentric liquid lens (108) based on a vertical distance between the rails (104) and the line scan camera (110) that is coupled to the UAV (102). This is because, the vertical distance between the rails (104) and the line scan camera (110) varies from time-to-time during inspection of defects in the rails (104). Such variations in the vertical distance between the rails (104) and the line scan camera (110) cause the line scan camera (110) to capture images including contrast issues.
[0096] For example, the vertical distance between the rails (104) and the line scan camera (110) will decrease when the UAV (102) carrying the line scan camera (110) moves downwards from an associated current position to a lowered position during inspection. In another example, the vertical distance between the rails (104) and the line scan camera (110) will increase when the UAV (102) carrying the line scan camera (110) moves upwards from a current position to a raised position. Generally, such downward and upwards movements of the UAV (102) introduce contrast issues in images captured by the line scan camera (110) due to variations that occur in the vertical distance between the rails (104) and the line scan camera (110) during the downwards and upwards movements of the UAV (102). The vertical distance between the rails (104) and the line scan camera (110) also varies depending on an inclination of a ground surface on which the rails (104) are laid. For example, the vertical distance between the line scan camera (110) and the rails (104) that are laid on an ascending ground surface will be comparatively less than the vertical distance between the line scan camera (110) and the rails (104) that are laid on a descending ground surface. This inclined nature of the ground surface on which the rails (104) are laid often causes the line scan camera (110) to capture images including contrast issues.
[0097] In order to address the aforementioned issues, the image quality management system (112) dynamically adjusts the lens parameters such as a lens focus or a working distance of the telecentric liquid lens (108) during an inspection based on the vertical distance between the rails (104) and the line scan camera (110). Dynamically adjusting the lens parameters during inspection based on the vertical distance suppresses contrast issues at an image acquisition level itself. Further, dynamically adjusting the lens and/or camera parameters enables the line scan camera (110) to consistently capture uniform intensity, high contrast, and optimally exposed images of the rails (104), which are suitable for identifying defects in the rails (104) accurately by the rail defect identification system (114).
[0098] According to aspects of present disclosure, the rail defect identification system (114) determines widths of switch and/or joint gaps in the rails (104) accurately when compared to a conventional inspection system that uses an area scan camera and a fixed focus lens. For example, the following table provides a corresponding accuracy of a width of a switch gap (802) determined by each of the rail defect identification system (RDIS) (114) and the conventional inspection system (CIS) from an exemplary image (804).

Table 1 – Accuracy of a switch gap width determined by CIS and RDIS (114)

Inspection System AW in millimeter(mm) NOPE NOPL NOAP DW Accuracy
CIS 3 mm 12 4 8 2 mm 66.67%
RDIS (114) 3 mm 50 0-4 46-50 2.76 - 3 mm 92 – 100%

[0099] In Table 1, ‘AW’ corresponds to an actual width of a switch gap (802) in the image (804), ‘NOPE’ corresponds to a number of pixels expected to be present in the switch gap (802) region, and ‘NOPL’ corresponds to a number of pixels lost due to erosion and/or dilation effects. Further, ‘NOAP’ corresponds to a number of actual pixels present in the switch gap (802) region, and ‘DW’ corresponds to a width of the switch gap (802) determined by the conventional inspection system or the rail defect identification system (114).
[00100] In case of conventional inspection system, an area scan camera captures the image (804) of the switch gap (802). As the area scan camera is generally known to output low pixel resolution images when compared to the line scan camera (110), the switch gap (802) of 3 mm in the image (804) includes 12 pixels. In other words, a width of each pixel in the switch gap (802) region may correspond to 0.25 mm, and therefore, the switch gap (802) of 3 mm includes 12 pixels ideally.
[00101] However, the image (804) of the rails (104) captured in a bright lighting condition by the area scan camera may include only 8 pixels in the switch gap (802) region. This is because some of the pixels, for example 4 out of 12 pixels, may be lost due to erosion effects when the image (804) is captured in the bright lighting conditions. When the conventional inspection system determines a width of the switch gap (802) from such an image (804) having only 8 pixels in the switch gap (802) region of 3 mm, the conventional inspection system will incorrectly determine the width of the switch gap (802) as 2 mm as there are only 8 pixels in the switch gap (802) region and the width of each pixel is 0.25 mm. This incorrect determination of the width of the switch gap (802) causes sharing of a false alert message to a railway track repair and maintenance division. Consequently, the railway track repair and maintenance division may unnecessarily send the service personnel to a geolocation where the switch gap (802) exists even though no actual defects exist in the railway track (106) in such a scenario.
[00102] In contrast to the conventional inspection system, the RTI system (100) uses the line scan camera (110) instead of the area scan camera for capturing the image (804) of the switch gap (802). The line scan camera (110) generally outputs high pixel resolution images. Hence, in the image (804) captured by the line scan camera (110), the switch gap (802) of 3 mm may include 50 pixels instead of 12 pixels as is the case in the conventional inspection system. In other words, a width of each pixel in the switch gap (802) region may correspond to 0.06 mm, and therefore, the switch gap (802) of 3 mm includes 50 pixels ideally.
[00103] Unlike the conventional inspection system in which the pixels may be lost due to erosion effects, the pixels in the image (804) captured by the line scan camera (110) will not be lost even when the image (804) is captured in the bright lighting condition. This is because the image quality management system (112) dynamically adjusts the lens parameters of the telecentric liquid lens (108) and/or the camera parameters of the line scan camera (110) according to the prevailing lighting condition prior to capturing of the image (804) by the line scan camera (110). Adjusting the lens and/or camera parameters prior to capturing of the image (804) by the line scan camera (110) prevents the erosion effects and preserves all pixels in the image (804). Hence, the switch gap (802) in the image (804) captured by the line scan camera (110) will be accurately illustrated using all 50 pixels without any loss. When the rail defect identification system (114) determines a width of the switch gap (802) from such an image (804) having all 50 pixels in the switch gap (802) region of 3 mm, the rail defect identification system (114) will accurately determine the width of the switch gap (802) as 3 mm. This is because, there are all 50 pixels exist in the switch gap (802) region of 3 mm and the width of each such pixel is 0.06 mm.
[00104] Additionally, the rail defect identification system (114) is still able to determine the width of the switch gap (802) accurately even if erosion effects were to be present. For example, the number of pixels that effectively exist in the switch gap (802) region in the image (804) captured by the line scan camera (110) will be 46 when 4 out of 50 pixels are lost due to erosion effects. When the rail defect identification system (114) determines the width of the switch gap (802) from such an image (804) having 46 pixels that exist in the switch gap (802) region of 3 mm, the rail defect identification system (114) will determine the width of the switch gap (802) as 2.76 mm as there are 46 pixels in the switch gap (802) region and the width of each pixel is 0.06 mm. It may be noted that the width of the switch gap (802) determined as 2.76 mm by the rail defect identification system (114) is much closer to the actual width of 3 mm when compared to the width of the switch gap (802) determined as 2 mm by the conventional inspection system. Thus, the RTI system (100), employing the line scan camera (110) and the telecentric liquid lens (108), enables the rail defect identification system (114) to determine widths of the switch and/or joint gaps in the rails (104) more accurately.
[00105] Referring back to FIG. 3, at step (310), the line scan camera (110) continuously captures the subsequent set of images of the rails (104) after the image quality management system (112) adjusts the one or more lens parameters and/or the one or more camera parameters for enabling the rail defect identification system (114) to identify defective switch and/or joint gaps if any in the rails (104) from the subsequent set of images. An exemplary methodology used to identify defective switch and joint gaps in the rails (104) is described subsequently in detail with reference to FIGS. 9A-B.
[00106] FIGS. 9A-B illustrate a flow diagram depicting an exemplary method (900) used by the rail defect identification system (114) to identify defective switch and joint gaps in the rails (104). At step (902), the rail defect identification system (114) combines the subsequent set of images of the rails (104) corresponding to line-by-line high contrast and optimally exposed images captured by the line scan camera (110) to generate a color image of a particular portion of the rails (104). For example, FIG. 10A illustrates an exemplary illustration depicting the color image (1000) of the particular portion of the rails (104) generated by the rail defect identification system (114) by combining the line-by-line high contrast and optimally exposed images of the rails (104) captured by the line scan camera (110).
[00107] At step (904), the rail defect identification system (114) converts the color image (1000) into a greyscale image. For example, FIG. 10B illustrates an exemplary illustration depicting the greyscale image (1002) that is derived from the image (1000). In certain embodiments, the rail defect identification system (114) identifies edges (1003A-B) of the rails (104) from the greyscale image (1002) post converting the image (1000) into the greyscale image (1002). For instance, the rail defect identification system (114) identifies the edges (1003A-B) of the rails (104) from the greyscale image (1002) using a canny edge detection technique.
[00108] Post identifying the edges (1003A-B) of the rails (104), at step (906), the rail defect identification system (114) performs image segmentation and converts the greyscale image (1002) including the edges (1003A-B) of the rails (104) into a binary image. In one embodiment, the rail defect identification system (114) uses filtering and morphological algorithm for performing image segmentation and converting the greyscale image (1002) into the binary image. For instance, FIG. 10C illustrates an exemplary illustration depicting the binary image (1004), including the edges (1003A-B) of the rails (104) that are generated by the rail defect identification system (114) from the greyscale image (1002) using filtering and morphological algorithm. It may be noted that image details such as the edges (1003A-B) of the rails (104) are clearly visible in FIG. 10B as FIG. 10B corresponds to the greyscale image (1002) that is derived from the color image (1000). However, the image details such as the edges (1003A-B) of the rails (104) are not clearly visible in FIG. 10C as FIG. 10C corresponds to the binary image (1004) that is obtained from the greyscale image (1002).
[00109] At step (908), the rail defect identification system (114) detects a first rail region including a joint gap and a second rail region including a switch gap from the binary image (1004) using a pattern matching technique. It may be noted that the binary image (1004) may not always include images of both the joint gap and the switch gap. The binary image (1004) may include only an image of the joint gap or only an image of the switch gap. However, for the sake of simplicity, the binary image (1004) is subsequently described to include images of both the joint and switch gaps. For example, FIG. 10D illustrates an exemplary illustration depicting a first rail region (1006) including a joint gap (1008) and a second rail region (1010) including a switch gap (1012) in the rails (104) that are detected from the binary image (1004) using the pattern matching technique by the rail defect identification system (114). At step (910), the rail defect identification system (114) crops an image portion including the first rail region (1006) from the greyscale image (1002).
[00110] Further, at step (912), the rail defect identification system (114) determines a width of the joint gap (1008) in the first rail region (1006) from a greyscale image (1014). In one embodiment, the rail defect identification system (114) determines the width of the joint gap (1008) in the first rail region (1006) using a line profile technique and a contour measurement technique. In particular, the rail defect identification system (114) generates a plurality of horizontal lines (1016A-C) across the joint gap (1008) using the line profile technique, as depicted in FIG. 10E. For simplicity, the rail defect identification system (114) is depicted and subsequently described to generate only three horizontal lines (1016A-C) across the joint gap (1008) for determining the width of the joint gap (1008). However, in practical scenarios, the rail defect identification system (114) may generate multiple such horizontal lines across the joint gap (1008) for accurately determining the width of the joint gap (1008).
[00111] Upon generating the horizontal lines (1016A-C), the rail defect identification system (114) identifies a left boundary (1018A) and a right boundary (1018B) of the joint gap (1008) lying in a first horizontal line (1016A). To that end, the rail defect identification system (114) continuously identifies a pixel intensity of each and every pixel in the first horizontal line (1016A) from a left-to-right direction (1020). When the rail defect identification system (114) continuously identifies the pixel intensity of each and every pixel in the first horizontal line (1016A) from the left-to-right direction (1020), the rail defect identification system (114) will encounter a particular pixel in the joint gap (1008) region whose pixel intensity will be equivalent to zero. The rail defect identification system (114) then identifies a specific pixel point in the first horizontal line (1016A) whose pixel intensity that is equivalent to zero as the left boundary (1018A) of the joint gap (1008) lying in the first horizontal line (1016A).
[00112] Similarly, the rail defect identification system (114) continuously identifies a pixel intensity of each and every pixel in the first horizontal line (1016A) from a right-to-left direction (1022). When the rail defect identification system (114) continuously identifies the pixel intensity of each and every pixel in the first horizontal line (1016A) from the right-to-left direction (1022), the rail defect identification system (114) will encounter a particular pixel in the joint gap (1008) region whose pixel intensity will be equivalent to zero. The rail defect identification system (114) then identifies a specific pixel point in the first horizontal line (1016A) whose pixel intensity that is equivalent to zero as the right boundary (1018B) of the joint gap (1008). Subsequently, the rail defect identification system (114) determines a number of pixels that exist in between the identified left and right boundaries (1018A-B) of the joint gap (1008), for example, using a contour measurement technique.
[00113] In one embodiment, the rail defect identification system (114) similarly determines left and right boundaries (1024A-B) of the joint gap (1008) lying in a second horizontal line (1016B), and left and right boundaries (1026A-B) of the joint gap (1008) lying in a third horizontal line (1016C). In addition, the rail defect identification system (114) determines a number of pixels in between the identified left and right boundaries (1024A-B) and a number of pixels in between the identified left and right boundaries (1026A-B). The rail defect identification system (114) then determines an average of the number of pixels determined in between each of the left and right boundaries (1018A-B) lying in the first horizontal line (1016A), the left and right boundaries (1024A-B) lying in the second horizontal line (1016B), and the left and right boundaries (1026A-B) lying in the third horizontal line (1016C). In certain embodiments, the average of number of pixels, thus determined by the rail defect identification system (114), corresponds to the width of the joint gap (1008) in the first rail region (1006).
[00114] In certain embodiments, at step (914), the rail defect identification system (114) crops an image portion including the second rail region (1010) from the image (1000) corresponding to the color image to generate another cropped image. For example, FIG. 11A illustrates an exemplary illustration depicting an image (1100) including the second rail region (1010) that is cropped from the image (1100). At step (916), the rail defect identification system (114) performs gap segmentation and converts the cropped image (1100) into a gap segmented image. For example, FIG. 11B illustrates an exemplary illustration depicting a gap segmented image (1102) that is generated from the cropped image (1100) using the rail defect identification system (114). At step (918), the rail defect identification system (114) converts the gap segmented image (1102) into another color image (1104) and determines a width of the switch gap (1012) in the second rail region (1010) from the color image (1104).
[00115] In one embodiment, the rail defect identification system (114) determines the width of the switch gap (1012) using the line profile and contour measurement techniques similar to the techniques that are previously described with respect to determining the width of the joint gap (1008) by the rail defect identification system (114). For example, the rail defect identification system (114) generates a plurality of horizontal lines such as a first horizontal line (1106A) and a second horizontal line (1106B) across the switch gap (1012). The rail defect identification system (114) then determines left and right boundaries (1108A-B) of the switch gap (1012) lying in the first horizonal line (1106A) and left and right boundaries (1110A-B) of the switch gap (1012) lying the second horizonal line (1106B) by analyzing intensities of pixels in the first and second horizontal lines (1106A-B), respectively. Subsequently, the rail defect identification system (114) determines a first number of pixels that exist in between the left and right boundaries (1108A-B) and a second number of pixels that exist in between the left and right boundaries (1110A-B). Further, the rail defect identification system (114) determines an average of the determined first number of pixels and the determined second number of pixels. In one embodiment, the average number of pixels, thus determined by the rail defect identification system (114), corresponds to the width of the switch gap (1012) in the second rail region (1010).
[00116] At step (920), the rail defect identification system (114) determines if the determined width of the joint gap (1008) is within a first designated range and if the determined width of the switch gap (1012) is within a second designated range. At step (922), the alerting system (126) transmits an alert message to the ground control station (120) when one or more of the determined width of the joint gap (1008) deviates from the first designated range and/or the determined width of the switch gap (1012) deviates from the second designated range. For example, the alerting system (126) transmits an alert message to the ground control station (120) when the determined width of the joint gap (1008) deviates from the first designated range. In one embodiment, the alert message transmitted to the ground control station (120) specifies the nature of the defect identified in the rails (104). In addition, the alert message also specifies the determined width of the defective joint gap (1008) and an associated geolocation for enabling the service personnel from the railway track repair and maintenance division to reach the particular geolocation and to carry out repair and/or maintenance activities.
[00117] In one embodiment, the alert message that is transmitted to the ground control station (120) corresponds to a text alert message, an audio alert message, or a video alert message. For example, the rail defect identification system (114) transmits the text alert message to the ground control station (120) that corresponds to a remote computer or a server upon identifying a defect in the rails (104). The text alert message, thus transmitted to the ground control station (120), may include an identified defect type that specifies if the identified defect type corresponds to a defective switch gap or a defective joint gap, and the width and the geolocation of the defective switch gap or joint gap identified in the rails (104). According to aspects of the present disclosure, the alerting system (126) includes a text-to-audio converter that converts the text message including the identified defect type, the width, and the geolocation of the defect identified in the rails (104) into an audio alert message. Further, the alerting system (126) transmits the audio alert message to the ground control station (120) instead of transmitting the alert message in a text format. Upon receiving such audio alert message, the ground control station (120) may provide, for example, a buzzer sound effect to alert a concerned person on the defect identified in the rails (104).
[00118] In certain embodiments, the alerting system (126) may also capture a video of a portion of the rails (104) including the defective switch or joint gap and transmit the video alert message to the ground control station (120). The video alert message, thus transmitted to the ground control station (120), may include the captured video, the width of the defective switch or joint gap, and the geolocation of the defective switch gap or joint gap identified in the rails (104).
[00119] According to aspects of the present disclosure, the alerting system (126) intelligently selects a recipient to whom the alert message is to be transmitted based on an amount of deviation between the width of the joint gap (1008) and the first designated range and/or an amount of deviation between the width of the switch gap (1012) and the second designated range. For example, the alerting system (126) transmits the alert message to the ground control station (120) maintained by a railway track maintenance division when the width of the joint gap (1008) or the width of the switch gap (1012) deviates up to a specific threshold of 10% from the first designated range and the second designated range, respectively. However, the alerting system (126) transmits the alert message to the ground control station (120) maintained by a railway track emergency division when the width of the joint gap (1008) or the width of the switch gap (1012) deviates by more than 10% from the first designated range and the second designated range, respectively. Any deviation that is greater than, for example, 10% may be considered as a serious rail defect, which could potentially cause trains to derail. Hence, the alerting system (126) transmits the alert message to the railway track emergency division when widths of the joint and/or switch gaps (1008 and 1012) deviate by more than 10% from their corresponding designated ranges for enabling the railway track emergency division to perform necessary corrective actions in an expedited manner and to save human lives.
[00120] As noted previously, conventional inspection systems use an area scan camera and a fixed focus lens for capturing images of the rails (104) and for identifying defects in the rails (104) from the captured images. However, the fixed focus lens causes the area scan camera to capture images that include parallax and aberration errors. The conventional inspection systems do no identify defects in the rails (104) accurately from such images including parallax and aberration errors. In contrast, the RTI system (100) described in the present disclosure uses the telecentric liquid lens (108) and the line scan camera (110). The RTI system (100) suitably adjusts the lens parameters of the telecentric liquid lens (108) using the associated image quality management system (112), which enables the line scan camera (110) to capture parallax and aberration errors free images of the rails (104). Hence, the RTI system (100) is able to identify defects in the rails (104) accurately from such error free images.
[00121] Further, the lens parameters of the fixed focus lens and the camera parameters of the area scan camera used in the conventional inspection systems are not alterable according to dynamically changing ambient lighting conditions. Hence, the images captured by the area scan camera in the conventional inspection systems are likely to be underexposed or overexposed images including contrast and saturation issues, which are not suitable for identifying defects in the rails (104) accurately. In contrast, the lens parameters of the telecentric liquid lens (108) and the camera parameters of the line scan camera (110) used in the RTI system (100) are alterable according to dynamically changing ambient lighting conditions. Hence, the telecentric liquid lens (108) and the line scan camera (110) enable the RTI system (100) to consistently capture uniform intensity, high contrast, and optimally exposed images, which are suitable for identifying defects in the rails (104) accurately.
[00122] In certain embodiments, the RTI system (100) employs a single telecentric liquid lens (108) and a single line scan camera (110) for capturing images of both the rails (104) as described previously with references to FIGS. 1-11C. Alternatively, the RTI system (100) may employ two separate units including a first unit that includes a first telecentric liquid lens and a first line scan camera, and a second unit that includes a second telecentric liquid lens and a second line scan camera. The first unit may capture images of one of the rails (104) and the second unit may separately and simultaneously capture images of another rail in the rails (104).
[00123] Although specific features of various embodiments of the present systems and methods may be shown in and/or described with respect to some drawings and not in others, this is for convenience only. It is to be understood that the described features, structures, and/or characteristics may be combined and/or used interchangeably in any suitable manner in the various embodiments shown in the different figures.
[00124] While only certain features of the present systems and methods have been illustrated and described herein, many modifications and changes will occur to those skilled in the art. It is, therefore, to be understood that the appended claims are intended to cover all such modifications and changes.

LIST OF NUMERAL REFERENCES:

100 Railway track inspection system
102, 604 Unmanned aerial vehicle
104 Rails
106 Railway track
108 Telecentric liquid lens
110 Line scan camera
112Image quality management system
114 Rail defect identification system
116 Location sensor
118 Power unit
120 Ground control station
122, 124 Communications link
126 Alerting system
202, 208 Rail switch and joint regions
204A-B Tongue tracks
206, 802, 1012 Switch gap
210, 1008 Joint gap
300-310 Steps of a method for capturing high contrast images of rails
402, 404 Horizontal axes
406, 408, 410, 416 Images related to parallax and aberration errors
412, 414 Regions of aberrated image
418 Corner of aberration free image
502 UAV position
504 Unmanned vehicle path center axis
506 Railway track center axis
508, 512, 606A-C Images of rail areas
510 UAV displaced position
514 UAV displacement distance
602 Area scan camera
608 Stitched image
610A-F Line-by-line rail images
700, 702, 704 Histograms
804 Rail image
900-922 Steps of a method for determining defects in rails
1000, 1100, 1104 Color images of rails
1002, 1014 Greyscale images of rails
1003A-B Edges of rails
1006, 1010 Rail regions
1016A-C, 1106A-B Horizontal lines across joint and switch gaps
1018A, 1024A, 1026A Left boundaries of joint gap
1020, 1022 Pixel analysis directions
1018B, 1024B, 1026B Right boundaries of joint gap
1102 Gap segmented image
1108A-B, 1110A-B Left and right boundaries of switch gap

, Claims:

We claim:

1. An inspection system (100), comprising:
a line scan camera (110) that comprises a single line of sensor pixels and is adapted to capture an initial line scan image of a target object (104) to be inspected;
an image quality management system (112) that is communicatively coupled to the line scan camera (110) and is adapted to determine a lighting condition prevailing in the surroundings of the target object (104);
a telecentric liquid lens (108) that is operatively coupled to the line scan camera (110) and the image quality management system (112), wherein the telecentric liquid lens (108) is adapted to adjust one or more associated lens parameters, the line scan camera (110) is adapted to adjust one or more associated camera parameters, or a combination thereof, based on one or more of the lighting condition determined by the image quality management system (112) and a vertical distance between the line scan camera (110) and the target object (104) for enabling the line scan camera (110) to capture a subsequent set of line scan images of the target object (104);
a defect identification system (114) that is operatively coupled to the line scan camera (110) and is adapted to process the subsequent set of line scan images of the target object (104) captured by the line scan camera (110) to identify if one or more defects exist in the target object (104); and
an alerting system (126) that is operatively coupled to the defect identification system (114) and is adapted to transmit an alert message to a ground control station (120) when the defect identification system (114) identifies the defects in the target object (104).

2. The inspection system (100) as claimed in claim 1, wherein the target object (104) corresponds to one or more rails (104) of a railway track (106), and wherein the defects identified in the target object (104) correspond to one or more of a defective joint gap and a defective switch gap in the rails (104) of the railway track (106).

3. The inspection system (100) as claimed in claim 2, wherein the inspection system (100) is operatively coupled to an unmanned aerial vehicle (102), wherein the unmanned aerial vehicle (102) moves over the railway track (106) such that an unmanned vehicle path center axis (504) aligns with a railway track center axis (506) during inspection of the defects in the rails (104) of the railway track (106).

4. The inspection system (100) as claimed in claim 2, wherein the alerting system (126) transmits the alert message to the ground control station (120) when the defect identification system (114) identifies one or more of the defective joint gap and the defective switch gap in the rails (104), wherein the alert message corresponds to one of a text alert message, an audio alert message, and a video alert message, and wherein the alert message specifies types of the defects comprising one or more of the defective joint gap and the defective switch gap identified in the rails (104), a width and a geographical location of the defective joint gap, and a width and a geolocation of the defective switch gap.

5. The inspection system (100) as claimed in claim 4, wherein the alerting system (126) transmits the alert message to the ground control station (120) for scheduled maintenance when the width of the defective joint gap or the width of the defective switch gap deviates from a corresponding designated range by less than a specific threshold, and
wherein the alerting system (126) transmits the alert message to the ground control station (120) for emergency maintenance when the width of the defective joint gap or the width of the defective switch gap deviates from the corresponding designated range by more than the specific threshold.

6. The inspection system (100) as claimed in claim 1, wherein the telecentric liquid lens (108) comprises one of an object space telecentric liquid lens, an image space telecentric liquid lens, a double telecentric liquid lens, bi-telecentric liquid lens, and a telecentric motorized lens, wherein the one or more associated lens parameters of the telecentric liquid lens (108) comprise one or more of a working distance, a lens focal axis, a lens orientation, a lens iris, a focal length, a lens zooming level, and a field of view of the telecentric liquid lens (108), wherein the one or more associated camera parameters of the line scan camera (110) comprise one or more of a shutter speed, gamma correction, a camera gain, white balance, and a color image format.

7. The inspection system (100) as claimed in claim 1, wherein the inspection system (100) comprises one or more of a road defect identifying system, a railway track inspection system, a manufacturing facility, a service station, an aerial photography system, a geographical mapping system, a package delivery system, an agriculture management system, a disaster management system, and a weather forecast system.

8. A method for inspecting a target object (104), comprising:
capturing an initial line scan image of the target object (104) using a line scan camera (110) that comprises a single line of sensor pixels;
determining a lighting condition prevailing in the surroundings of the target object (104) from the initial line scan image of the target object (104) by an image quality management system (112) that is communicatively coupled to the line scan camera (110);
adjusting one or more of one or more lens parameters of a telecentric liquid lens (108) and one or more camera parameters of the line scan camera (110) based on one or more of the lighting condition determined by the image quality management system (112) and a vertical distance between the line scan camera (110) and the target object (104);
capturing a subsequent set of line scan images of the target object (104) by the line scan camera (110) after adjusting one or more of the one or more lens parameters of the telecentric liquid lens (108) and the one or more camera parameters of the line scan camera (110);
processing the subsequent set of line scan images of the target object (104) captured by the line scan camera (110) by a defect identification system (114) to identify if one or more defects exist in the target object (104); and
transmitting an alert message to a ground control station (120) by an alerting system (126) when the defect identification system (114) identifies the defects in the target object (104).

9. The method claimed in claim 8, wherein the target object (104) corresponds to one or more rails (104) of a railway track (106), and wherein the defects identified in the target object (104) corresponds to one or more of a defective joint gap and a defective switch gap in the rails (104) of the railway track (106).

10. The method as claimed in claim 9, wherein adjusting one or more lens parameters of the telecentric liquid lens (108) comprises adjusting one or more of a focus and a working distance of the telecentric liquid lens (108) based on the vertical distance between the line scan camera (110) and the rails (104) when the initial line scan image of the target object (104) comprises one or more contrast issues.

11. The method as claimed in claim 9, wherein adjusting one or more of the one or more lens parameters of the telecentric liquid lens (108) and one or more camera parameters of the line scan camera (110) based on the determined lighting condition comprises:
identifying that the initial line scan image of the rails (104) captured by the line scan camera (110) is an underexposed image when pixel intensity values of a designated percentage of pixels in the initial line scan image fall in a first pixel intensity range;
identifying that the initial line scan image of the rails (104) captured by the line scan camera (110) is an overexposed image when pixel intensity values of the designated percentage of pixels in the initial line scan image fall in a second pixel intensity range;
performing a first corrective action when the initial line scan image of the rails (104) captured by the line scan camera (110) corresponds to the underexposed image, wherein the first corrective action comprises one or more of increasing a shutter speed of the line scan camera (110), increasing a brightness level of the line scan camera (110), increasing an iris value of the telecentric liquid lens (108), and increasing a gamma level of the line scan camera (110); and
performing a second corrective action when the initial line scan image of the rails (104) captured by the line scan camera (110) corresponds to the overexposed image, wherein the second corrective action comprises one or more of decreasing a shutter speed of the line scan camera (110), decreasing a brightness level of the line scan camera (110), decreasing an iris value of the telecentric liquid lens (108), and decreasing a gamma level of the line scan camera (110).

12. The method as claimed in claim 9, wherein capturing the initial line scan image of the rails (104) using the line scan camera (110) comprises adjusting a position of a mechanical aperture stop in the telecentric liquid lens (108) and a position of one or more of an imaginary entrance pupil and an imaginary exit pupil associated with the telecentric liquid lens (108) when the initial line scan image of the rails (104) captured by the line scan camera (110) comprises one or more of a parallax error and an aberration error.

13. The method as claimed in claim 12, wherein identifying if one or more defects exist in the rails (104) by the inspection system (100) comprises:
combining the subsequent set of line scan images of the rails (104) captured by the line scan camera (110) into an image (1000) of a particular portion of the rails (104);
converting the image (1000) into a greyscale image (1002);
performing image segmentation and converting the greyscale image comprising edges (1003A-B) of the rails (104) into a binary image (1004);
detecting a first rail region (1006) in the binary image (1004) comprising a joint gap (1008) and a second rail region (1010) in the binary image (1004) comprising a switch gap (1012) using a pattern matching technique;
cropping an image portion comprising the first rail region (1006) from the greyscale image (1002);
generating a first horizontal line (1016A) and a second horizontal line (1016B) across the joint gap (1008) in the first rail region (1006);
identifying a first left boundary (1018A) and a first right boundary (1018B) of the joint gap (1008) lying in the first horizontal line (1016A) and a first number of pixels that exist between the first left boundary (1018A) and the second right boundary (1018B);
identifying a second left boundary (1024A) and a second right boundary (1024B) of the joint gap (1008) lying in the second horizontal line (1016B) and a second number of pixels that exist between the second left boundary (1024A) and the second right boundary (1018B);
determining an average of the identified first number of pixels and the identified second number of pixels, wherein the determined average corresponds to a width of the joint gap (1008) in the first rail region (1006); and
identifying that the joint gap (1008) in the first rail region (1006) as the defective joint gap when the determined width of the joint gap (1008) deviates from a first designated range.

14. The method as claimed in claim 13, wherein identifying if one or more defects exist in the rails (104) by the inspection system (100) comprises:
cropping an image portion comprising the second rail region (1010) from the image (1000) to obtain a cropped image (1100);
performing a gap segmentation on the cropped image (1100) to obtain a gap segmented image (1102) from the cropped image (1100) and converting the gap segmented image (1102) into a color image (1104);
generating a first horizontal line (1106A) and a second horizontal line (1106B) across the switch gap (1012) in the second rail region (1010);
identifying a first left boundary (1108A) and a first right boundary (1108B) of the switch gap (1012) lying in the first horizontal line (1106A) and a first number of pixels that exist between the first left boundary (1108A) and the second right boundary (1108B);
identifying a second left boundary (1110A) and a second right boundary (1110B) of the switch gap (1012) lying in the second horizontal line (1106B) and a second number of pixels that exist between the second left boundary (1110A) and the second right boundary (1110B);
determining an average of the identified first number of pixels and the identified second number of pixels, wherein the determined average corresponds to a width of the switch gap (1012) in the second rail region (1010); and
identifying that the switch gap (1012) in the second rail region (1010) as the defective switch gap when the determined width of the switch gap (1012) deviates from a second designated range.

15. The method as claimed in claim 14, wherein the alert message that is transmitted to the ground control station (120) comprises the defects identified in the rails (104) comprising one or more of the defective joint gap and the defective switch gap, widths of one or more of the defective joint gap and the defective switch gap, and geographical locations of one or more of the defective joint gap and the defective switch gap.