DESC:RELATED ART

[0001] Embodiments of the present specification relate generally to inspecting integrity of seals, and more particularly to a seal inspection system and an associated method for inspecting integrity of container seals during movement of the containers on a packaging line.
[0002] In packaging industry, proper sealing of containers that are used for storing products such as food items, beverages, coffee powders, tea powders, and medicinal products is critical for maintaining the freshness and safety of products for a prolonged period of time. Any defects that occur during sealing of the containers may introduce moisture or other contaminants into the containers. The defective sealing, thus, results in loss of quality of the products and poses a health hazard to people consuming spoiled products, which further affects company’s reputation and brand name.
[0003] Accordingly, certain conventional seal inspection systems employ human resources to manually inspect seal quality and integrity of selected sample containers and identify if the seals contain any defects such as absence of seal, seals that are loosely fitted to the containers, and seals that include holes. However, manually inspecting seal quality and integrity requires a lot of time, cost, and human effort.
[0004] Accordingly, certain other seal inspection systems were developed to automatically inspect integrity of the seals. For example, Indian patent application No. 201721007306 describes one such seal inspection system that automatically detects defects in bottle seals. Specifically, the seal inspection system described in Indian patent application No. 201721007306 appears to include an induction sealing machine, and an accumulator conveyor that moves sealed bottles at a particular speed such that every single sealed bottle is positioned underneath a thermal image camera one after the other at a time interval of 12.5 milliseconds. Further, the thermal image camera is pre-calibrated to capture one image every 12.5 milliseconds to capture corresponding images that are processed to determine if the bottles are sealed properly.
[0005] However, the seal inspection system described in Indian patent application No. 201721007306 requires frequent calibration of the thermal image camera to match number of bottles to be sealed at a particular instant of time. For example, the thermal image camera needs to be re-calibrated to capture one image every 6 milliseconds when the accumulator conveyor may be reconfigured to move sealed bottles at an increased speed such that the sealed bottles are placed underneath the thermal image camera one after other at a time interval of 6 milliseconds to double the sealing capacity.
[0006] Further, the sealing machine described in the previously noted Indian patent application includes a stopper mechanism that controls operations of the accumulator conveyor to stop the movement of the bottles every few milliseconds. Specifically, the stopper mechanism controls operations of the accumulator conveyor such that the thermal image camera can capture images of the bottles when the bottles are in a stationary condition to purportedly allow capture of clearer images. In addition, the sealing machine includes a sensor that determines if the bottles are in the stationary condition before capturing images of the bottles. Stopping the movement of the bottles and then capturing images of the bottle affects productivity and reduces a number of bottles sealed over a designated period of time. Further, addition of such a stopper mechanism adds complexity and increases an overall cost of the sealing machine.
[0007] Accordingly, there remains a need for an improved and cost-effective seal inspection system that automatically inspects the integrity of seals accurately, while also allowing for easy reconfiguration of the sealing mechanism for enhancing productivity.

BRIEF DESCRIPTION

[0008] It is an objective of the present disclosure to provide a seal inspection system. The seal inspection system includes a first imaging sensor having a first type that captures a first video of a first container and a second imaging sensor having a second type different from the first type that captures a second video of the first container while a sealing operation is being performed by a sealing machine on a second container. The first video and the second video are captured by the first imaging sensor and the second imaging sensor simultaneously and at the same frame rate when the first container is positioned in a stationary condition. The seal inspection system further includes a frame selection system, a frame processing system, and a defect classification system operatively coupled to the first imaging sensor and the second imaging sensor. The frame selection system identifies a first subset of frames in the first video whose associated blur scores are greater than a designated blur score, and identifies a second subset of frames in the second video corresponding to the first subset of frames.
[0009] The frame processing system processes a selected frame from the second subset of frames to identify a hot region that corresponds to a sealing region of the container. The defect classification system classifies the hot region based on different patterns of seals learnt during a training stage of the defect classifying system to identify if the first container includes a defective seal. An alerting system communicatively coupled to the defect classification system and adapted to generate an alert upon identifying that the first container includes a defective seal. The first imaging sensor corresponds to a red-green-blue camera. The second imaging sensor corresponds to a thermographic camera. The first container corresponds to a sealable container for storing one or more of a food item, beverage, coffee powder, tea powder, a chemical product, an industrial product, an aerosol, and a medicinal product. The seal inspection system includes a defective container discarding system that automatically removes the first container identified to include the defective seal from the sealing machine.
[0010] The defective container discarding system includes a defective container picking unit controller, a set of motors, a set of defective container picking units, and a defective container collection tray. The set of defective container picking units includes one or more of a robotic arm and a vacuum tube adapted to automatically move the first container identified to include the defective seal from the sealing machine to the defective container collection tray. The alerting system includes one or more of a display device and an audio-visual system. The alerting system presents one or more of an audio and a video notification to alert an operator of the sealing machine regarding a count of defective containers, a position of one or more of the defective containers, a type of the identified defect, and an issue with the sealing machine.
[0011] It is another objective of the present disclosure to provide a method for inspecting a seal of a container. The method includes capturing a first video of a first container by a first imaging sensor having a first type and a second video of the first container by a second imaging sensor having a second type different from the first type while a sealing operation is being performed by a sealing machine on a second container. The first video and the second video are captured by the first imaging sensor and the second imaging sensor simultaneously at the same frame rate when the first container is positioned in a stationary condition. The method further includes determining a corresponding blur score for each frame in the first video and identifying a first subset of frames in the first video whose associated blur scores are greater than a designated blur score, and identifying a second subset of frames in the second video corresponding to the first subset of frames and selecting a frame from the identified second subset of frames.
[0012] Furthermore, the method includes identifying a hot region in the selected frame that corresponds to a sealing region of the container, and classifying the hot region based on different patterns of seals learnt during a training stage of an associated defect classifying system to identify if the first container includes a defective seal. Moreover, the method includes generating an alert by an alerting system upon identifying that the first container includes a defective seal. Identifying a second subset of frames in the second video includes determining unique identifiers of the first subset of frames. The determined unique identifiers correspond to sequence numbers of the first subset of frames in the first video. Identifying the second subset of frames in the second video further includes selecting frames from the second video that correspond to the sequence numbers of the first subset of frames to identify the second subset of frames.
[0013] Identifying the hot region in the selected frame includes interpolating a corresponding temperature value for each pixel in the selected frame using a stored temperature-pixel intensity value correlation based on a maximum temperature value in the selected frame, a minimum temperature value in the selected frame, a first pixel intensity threshold corresponding to the maximum temperature value, and a second pixel intensity corresponding to the minimum temperature value. Further, the method includes segregating a set of pixels in the selected frame whose associated temperature values are greater than a designated temperature threshold as the hot region, and cropping the set of segregated pixels indicative of the hot region from the selected frame. Furthermore, the method includes converting a cropped portion including the set of segregated pixels into a binary image frame and sharing the binary image frame with the defect classification system. Moreover, the method includes training the defect classification system to learn one or more patterns indicative of one or more types of defective seals. The one or more types of defective seals includes an incomplete seal and a seal including non-uniform thickness. The defect classification system learns one or more patterns indicative of incomplete seals using a first set of reference binary image frames, and learns one or more patterns indicative of one or more of complete seals using a second set of reference binary image frames.
[0014] Further, the method includes automatically moving the first container identified to include the defective seal from the sealing machine to a defective container collection tray by a defective container discarding system. Furthermore, the method includes pre-calibrating the defective container discarding system to store specific numbers of motor revolutions required to move an associated defective container picking unit from a default position to different positions in the moveable tray in the sealing machine and from the different positions in the moveable tray to the defective container collection tray. Moreover, the method includes rotating the motor in a first direction by a stored specific number of motor revolutions to position the defective container picking unit in an identified position adjacent to the first container identified to include a defective seal. In addition, the method includes removing the first container from the sealing machine by the associated defective container picking unit. Further, the method includes rotating the motor in a second direction opposite to the first direction by a stored specific number of motor revolutions such that the associated defective container picking unit moves the first container into the defective container collection tray.

BRIEF DESCRIPTION OF DRAWINGS

[0015] 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:
[0016] FIG. 1 illustrates a block diagram depicting an exemplary seal inspection system that is operatively coupled to a sealing machine for inspecting integrity of seals that are added to one or more containers by the sealing machine, in accordance with aspects of the present disclosure;
[0017] FIGS. 2A-B illustrate a flowchart depicting an exemplary method for identifying if one or more of the containers include a defective seal using the seal inspection system of FIG. 1, in accordance with aspects of the present disclosure;
[0018] FIGS. 3A-B illustrate a flowchart depicting an exemplary method for selecting and processing a suitable frame from a video captured by the seal inspection system of FIG. 1 for identifying if one or more of the containers include a defective seal, in accordance with aspects of the present disclosure;
[0019] FIG. 4A illustrates an exemplary illustration of the selected frame that is processed by the seal inspection system of FIG. 1 to identify if one or more of the containers include a defective seal, in accordance with aspects of the present disclosure;
[0020] FIG. 4B illustrates an exemplary image generated by cropping the selected frame of FIG. 4A, in accordance with aspects of the present disclosure;
[0021] FIG. 4C illustrates an exemplary binary image frame that is obtained from the cropped image of FIG. 4B, in accordance with aspects of the present disclosure;
[0022] FIG. 4D illustrates an exemplary image of a sealing region of a container that completely lacks a seal, in accordance with aspects of the present disclosure;
[0023] FIGS. 5A-F illustrate exemplary binary images depicting certain exemplary sealing regions of containers including defective seals, in accordance with aspects of the present disclosure;
[0024] FIGS. 5G-L illustrate exemplary binary images depicting sealing regions of one or more containers including complete and good quality seals, in accordance with aspects of the present disclosure; and
[0025] FIG. 6 illustrates a schematic diagram depicting an exemplary defective container discarding system that moves one or more of the containers identified to include a defective seal from the sealing machine of FIG. 1 to one or more defective container collection trays, in accordance with aspects of the present disclosure.

DETAILED DESCRIPTION

[0026] The following description presents an exemplary system that automatically inspects integrity of seals. Particularly, embodiments described herein disclose a seal inspection system that identifies containers including defective seals such as containers that include only partial seals and containers that lack seals completely. The seal inspection system further includes a defective container discarding system that automatically removes containers including such defective seals from a sealing machine to ensure such defective containers do not reach end customers.
[0027] As noted previously, inspecting seal quality and integrity is a critical step in the production environment to ensure quality and safety of contents packaged in the containers. Certain conventional seal inspection systems include a thermal image camera to capture images of sealed containers that are processed to verify seal integrity. Additionally, conventional seal inspection systems include a stopper mechanism that is generally programmed to stop sealing operations performed by the sealing machine, for example, for a few milliseconds per container. The stopper mechanism, thereby, is supposed to enable an associated thermal image camera to capture potentially clearer images of the container in a stationary condition post sealing to inspect quality and integrity of seals from the captured images.
[0028] However, stopping sealing operations performed by the sealing machine for even a few milliseconds affects productivity, thus significantly reducing a total number of containers sealed over a particular period of time. Further, the stopper mechanism and any sensor mechanism conventionally added to the sealing machine to stop the containers in a stand-still position and to verify if the containers are actually in the stand-still condition, respectively, are complex and expensive, and thereby increase an overall cost of the sealing machine.
[0029] Such conventional seal inspection systems that rely on the stopper mechanism to stop the sealing operations and image sealed containers in order to capture clear images. In contrast, the seal inspection system described in the present disclosure employs a thermographic camera that continuously captures images of sealed containers as a video without needing to stop the sealing operation performed by a sealing machine. Further, the seal inspection system processes the continuously captured video to identify integrity of the seals formed on the containers. The continuous imaging, thus, allows the present seal inspection system to achieve a significantly faster rate at which the containers are sealed. Additionally, the present seal inspection system eliminates the necessity of employing a conventional stopper mechanism and a sensing mechanism for detecting if a container is in stand-still condition that are complex and expensive.
[0030] It may be noted that different embodiments of the present seal inspection system may be used to inspect integrity and quality of different types of sealed containers. For example, the seal inspection system may be used to inspect integrity and quality of sealed plastic bottles containing chemicals or beverages, and sealed containers storing food items, medicinal products, industrial products, aerosols, and/or kitchen ingredients such as coffee powders and tea powders. However, for clarity, an embodiment of the seal inspection system is described in FIG. 1 in greater detail with reference to inspecting integrity and quality of sealed containers storing coffee powder.
[0031] FIG. 1 illustrates a block diagram depicting an exemplary seal inspection system (100) that is operatively coupled to a sealing machine (102) for inspecting integrity of seals that are added to one or more containers (104A-J) by the sealing machine (102). In one embodiment, the sealing machine (102) seals the containers (104A-J) in an airtight manner to preserve products stored within the containers (104A-J). For sealing the containers (104A-J) hermetically, the sealing machine (102) includes a set of heat pistons (106A-E) and a moveable tray (108).
[0032] In certain embodiments, the moveable tray (108) includes one or more rows of openings for accommodating the containers (104A-J) that are to be sealed. For example, the moveable tray (108) includes a first row of openings (110) that accommodate a first set of containers (104A-E) and a second row of openings (112) that accommodate another set of containers (104F-J) that are already sealed. For sealing the first set of containers (104A-E), in one embodiment, the sealing machine (102) aligns the first set of containers (104A-E) including covers (105A-E) that act as closures of the first set of containers (104A-E) below the heat pistons (106A-E). Subsequently, the sealing machine (102) moves the heat pistons (106A-E) downward to a designated position and configures the heat pistons (106A-E) to activate and apply heat on the covers (105A-E) such that the covers (105A-E) fasten to designated sealing regions (114A-E) of the corresponding containers (104A-E). Thus, the sealing machine (102) seals the containers (104A-E) with the corresponding covers (105A-E).
[0033] After sealing the containers (104A-E), the sealing machine (102) moves the moveable tray (108), for example, in a horizontal direction (116) such that the sealed containers (104A-E) are moved to a position corresponding to the second row of openings (112). Subsequently, the sealing machine (102) positions a second set of containers (104K-O) underneath the heat pistons (106A-E) for enabling the heat pistons (106A-E) to heat seal the second set of containers (104K-O).
[0034] In certain scenarios, the sealing machine (102) may fail to seal the containers (104A-E) adequately, thereby causing one or more sealing defects in one or more of the containers (104A-E). The resulting sealing defects, for example, may include an undesirably narrow seal, irregular sealing cross section leading to wrinkles, shrinkage, and deformation of seal, and contamination of the sealing regions (114A-E) by foreign matter leading to sub-optimal seals.
[0035] For example, the heat pistons (106A-E) may fail to heat the covers (105A-E) and/or the sealing regions (114A-E) of the containers (104A-E) uniformly and sufficiently, which leads to formation of defective seals. Such a defect may occur when the container (104A) positioned below the heat piston (106A) is not properly aligned with the heat piston (106A). Consequently, the heat piston (106A) comes into contact with only a small portion of the entire sealing region (114A) and/or cover (105A) of the container (106A), and thus, applies heat only to the small portion of the sealing region (114A) and/or cover (105A), while the other portions remain unheated. Further, in this example, the cover (105A) may fasten only to the small portion of the entire sealing region (114A) that is heated by the heat piston (106A), while failing to adhere to the rest of the unheated portion, thereby leading to only a partial closure or sealing of the container (104A).
[0036] In another example, impurities deposited on the sealing region (114A) or the cover (105A) and/or impurities deposited on the heat piston (106A) on one or more of associated exterior surfaces prevent transmission of heat. Presence of impurities, thus, fails to allow the heat piston (106A) to come into contact with the entire sealing region (114A) and/or the cover (105A), thereby failing to uniformly and adequately heating the sealing region (114A) and/or the cover (105A). As a result, the cover (105A) may fail to affix to the sealing region (114A) of the container (104) properly, which causes the container (104A) to remain open partially or completely, thus hindering quality of the associated contents, and in turn, endangering life and safety of end consumers.
[0037] In order to automatically identify one or more of such containers (104A-E) that do not include proper seals, the seal inspection system (100) includes a first imaging sensor (118) that is of a first type, a second imaging sensor (120) that is of a second type and is different from the first imaging sensor (118). An example of the first imaging sensor (118) includes a red-green-blue (RGB) camera (118). An example of the second imaging sensor (120) includes a thermographic camera (120).
[0038] In one embodiment, both the RGB camera (118) and the thermographic camera (120) simultaneously and independently capture a first RGB video and a second thermographic video, respectively of different operations performed by the sealing machine (102) during sealing of the containers (104A-E) in order to identify a subset of the containers (104A-E) including partial or no seals.
[0039] For example, the RGB camera (118) and the thermographic camera (120) capture the first RGB video and the second thermographic video including a first operation performed by the sealing machine (102). The first operation, for example, entails heating of the sealing regions (114A-E) and the covers (105A-E) of a first set of the containers (104A-E) by the heat pistons (106A-E). Further, the first RGB video and the second thermographic video captured by the RGB camera (118) and the thermographic camera (120) may additionally include a second operation performed by the sealing machine (102). The second operation, for example, includes transporting the first set of containers (104A-E) to a region corresponding to the adjacent row of openings (112). In addition, the first RGB video and the second thermographic video may also include a third operation performed by the sealing machine (102). The third operation, for example, includes positioning the first set of containers (104A-E) in the region corresponding to the adjacent row of openings (112) in a stationary condition for a designated period during which the heat pistons (106A-E) heat sealing regions (114K-O) and covers (105K-O) of a second set of containers (104K-O).
[0040] In order to accurately identify one or more of the first set of containers (104A-E) including only partial seals or no seals, the seal inspection system (100) identifies a specific sub-portion in the second thermographic video that includes images of the first set of containers (104A-E) that were captured when the first set of containers (104A-E) were positioned in the stationary condition for better clarity.
[0041] To that end, the seal inspection system (100) includes a frame selection system (122). Generally, images of the first set of containers (104A-E) that were captured while the first set of containers (104A-E) was moving on the moveable tray (108) include blur-related errors. Inspecting the integrity of seals added to the first set of containers (104A-E) from such blurred images leads to inaccurate identification of one or more of the first set of containers (104A) including a defective seal. In order to address the aforementioned issue, the frame selection system (122) identifies blur-free image frames in the second thermographic video that were captured when the first set of containers (104A-E) is positioned in the stationary condition, and uses only such blur-free image frames for inspecting the seal quality and integrity added to the first set of containers (104A-E). To that end, the frame selection system (122) determines a corresponding blur score for each image in the first RGB video that is captured by the RGB camera (118). Subsequently, the frame selection system (122) identifies a first subset of images from the first RGB video that is captured by the RGB camera (118) whose associated blur scores are greater than a designated blur score.
[0042] In certain embodiments, associated blur scores being greater than the designated blur score indicates that the first subset of images are blur-free images, for example, captured by the RGB camera (118) when the first set of containers (104A-N) are held in the stationary condition. Subsequently, the frame selection system (122) identifies a second subset of images that are blur-free frames in the second thermographic video captured by the thermographic camera (120). For example, the frame selection system (122) selects image frames from the second thermographic video that correspond to the first subset of images using corresponding unique identifiers, as described in detail with reference to FIGS. 2A-B. In certain embodiments, inspecting the integrity of seals added to the first set of containers (104A-E) from the second subset of images that are blur-free improves an accuracy with which one or more of the first set of containers (104A-E) including a defective seal is identified.
[0043] The second subset of images, thus identified by the frame selection system (122), correspond to the images of the first set of containers (104A-E) that are captured by the thermographic camera (120) when the first set of containers (104A-E) are held in the stationary condition. In one embodiment, the frame selection system (122) selects any one image from the identified second subset of images, and further provides the selected image as an input to a frame processing system (124) in the seal inspection system (100).
[0044] Subsequently, the frame processing system (124) processes the selected image, as described in detail with reference to FIGS. 3A-B, and provides the processed image as an input to a defect classifying system (126) in the seal inspection system (100). Subsequently, the defect classifying system (126) uses the processed images to classify each of the first set of containers (104A-E) either as a non-defective container with no defective seal or as a defective container with partial or no seal based on a training previously provided to the defect classifying system (126), as described in detail with reference to FIGS. 3A-B.
[0045] To that end, in certain embodiments, the seal inspection system (100) and associated frame selection system (122), frame processing system (124), and defect classifying system (126) may be implemented either as independent units or as a combined image processing system by suitable code on a processor-based system, such as a general-purpose or a special-purpose computer. Accordingly, the seal inspection system (100) and associated frame selection system (122), frame processing system (124), and defect classifying system (126), for example, include one or more general-purpose processors, specialized processors, graphical processing units, microprocessors, programming logic arrays, field programming gate arrays, integrated circuits, systems on chips, and/or other suitable computing devices.
[0046] According to aspects of the present disclosure, one or more of the first set of containers (104A-E) identified to be defective by the defect classifying system (126) are picked from the moveable tray (108) to prevent the defective containers from proceeding, for example, to an associated assembly line/packaging conveyor (not shown in FIGS. 1-6). To that end, in one embodiment, the seal inspection system (100) further includes a defective container discarding system (128) that picks containers classified to be defective by the defect classifying system (126) from the moveable tray (108). Further, the defective container discarding system (128) moves the containers picked from the moveable tray (108), for example, to one or more defective container collection trays in the sealing machine (102), as described in detail with reference to FIG. 6. In addition, the seal inspection system (100) includes an alerting system (130) such as an audio-visual system and/or a display device. The alerting system (130) presents one or more of an audio and a video notification to alert an operator of the sealing machine (102) regarding a count of defective containers, a position of one or more of the defective containers in the sealing machine (102), types of defects such as partial seals, no seals, and/or non-uniform thickness seals identified in the defective containers, and issues identified with reference to the sealing machine (102). An exemplary method by which the seal inspection system (100) identifies if the containers (104A-E) include defective seals to discard the defective containers from further packaging is described subsequently with reference to FIGS. 2A-B.
[0047] FIGS. 2A-B illustrate a flowchart depicting an exemplary method (200) for identifying if one or more of the first set of containers (104A-E) include a defective seal using the seal inspection system (100). The order in which the exemplary method (200) 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.
[0048] As described with reference to in FIG. 1, the present seal inspection system (100) simultaneously performs both sealing and imaging operations. Particularly, after heat sealing the first set of containers (104A-E), a motorized conveyor (not shown in FIGS) in the sealing machine (102) moves and places the first set of containers (104A-E) in a stationary condition in a position corresponding to the second row of openings (112) while the second set of containers (104K-O) are moved and are positioned underneath the heat pistons (106A-E) for heat sealing of the second set of containers (104K-O).
[0049] Subsequently, at step (202), the RGB camera (118) captures a first RGB video and simultaneously the thermographic camera (120) captures a second thermographic video of the first set of containers (104A-E) positioned in the stationary condition. In particular, the first RGB video and the second thermographic video are simultaneously captured at the same frame rate while a sealing operation is being performed by the sealing machine (102) on a second set of containers (104K-O). In conventional seal inspection systems, a sealing machine suspends the sealing operation, for example, every few milliseconds for a designated period of time in order to enable a thermal image camera to capture images of sealed containers in a stationary condition. Suspending the sealing operation performed by the sealing machine once every few milliseconds affects productivity of container packaging operations.
[0050] However, the present seal inspection system (100) prevents a need to suspend the sealing operation while the thermographic camera (120) is capturing images of the containers (104A-J). Instead, in one embodiment, the thermographic camera (120) continuously captures images of the first set of containers (104A-E) that are positioned in the second row of openings (112), while the sealing machine (102) simultaneously heats sealing regions of the second set of containers (104K-O) that are positioned underneath the heat pistons (106A-E). Thus, both heating the sealing regions of the second set of containers (104K-O) and capturing images of the first set of containers (104A-E) that have previously been sealed are performed simultaneously without needing to stop the sealing operation performed by the sealing machine (102), which increases the productivity and a number of containers sealed over a particular period of time.
[0051] Further, in conventional seal inspection systems, a thermal image camera is pre-calibrated to capture one image of a container, for example, every 12 milliseconds. In order to double the sealing capacity, a conveyor in a conventional sealing machine moves sealed containers at an increased speed such that the sealed containers are placed underneath the thermal image camera in a stationary condition one after the other at a time interval of 6 milliseconds. Accordingly, in conventional systems, the thermal image camera needs to be re-calibrated to capture one image every 6 milliseconds.
[0052] However, the thermographic camera (120) used in the seal inspection system (100) does not require any such recalibration even when the sealing capacity needs to be doubled or modified. This is because the present thermographic camera (120) is configured to continuously capture the second thermographic video including images of the first set of containers (104A-E) without needing to change a rate of imaging to adjust to any change in sealing capacity.
[0053] At step (204), the frame selection system (122) determines a corresponding blur score for each frame in the first RGB video. Generally, frames of the first set of containers (104A-E) that are captured when the first set of containers (104A-E) are moving along with the moveable tray (108) include blurriness. In order to capture blur-free frames of containers, conventional seal inspection systems stop the movement of the containers and then capture image frames of the containers. However, stopping the movement of the containers and then capturing the image frames do not guarantee that the captured frames would be blur-free unless a thermal image camera is precisely timed to capture the image frames after the containers come to a stand-still position. Additionally, conventional seal inspection systems need to stop the sealing operation when stopping the movement of the containers in order to capture blur-free image frames. Unlike such conventional seal inspection systems, the present seal inspection system (100) uses the first image sensor (118) and the second image sensor (120) to continuously image the first set of containers (104A-E) positioned in a stationary condition to generate the first RGB video and the second thermographic video, respectively while the second set of containers (104K-O) is being simultaneously sealed by the sealing machine (102). The seal inspection system (100) determines blur scores for all frames in the first RGB video, selects a subset of frames from the first RGB video that are blur-free frames, and uses a second subset of frames identified from the second thermographic video corresponding to the subset of frames selected from the first RGB video for inspecting the seal integrity and quality. Using blur-free frames for seal inspection improves speed and accuracy with which the first set of containers (104A-E) including defective seals are detected.
[0054] In one embodiment, the frame selection system (122) determines the corresponding blur score for each frame in the first RGB video using an Open Computer Vision (OpenCV) technique such as a fast Fourier transform technique. For example, the first RGB video of the first set of containers (104A-E) captured while the sealing operation is being performed by the sealing machine (102) includes 30 frames. In this example, the frame selection system (122) determines a blur score for each of the 30 frames. An exemplary blur score determined for each of the 30 frames is provided subsequently in Table 1.
[0055] Table 1 – Exemplary blur scores of frames in the first RGB video

F1 … F10 F11 … F15 F16 … F20 F21 … F30
BS=2 … BS=4 BS=5 … BS=6 BS=12 … BS=13 BS=5 … BS=3

[0056] In Table 1, F1 to F30 correspond to frames 1 to 30, respectively in the first RGB video, and ‘BS’ corresponds to exemplary blur scores determined for the frames 1 to 30. Further, in the example noted previously, frames 1 to 10 capture a first sub-operation performed by the sealing machine (102) including heating the sealing regions (114A-E) and the covers (105A-E) of the first set of containers (104A-E) using the heat pistons (106A-E). Frames 11 to 15 capture a second sub-operation performed by the sealing machine (102) including moving the first set of containers (104A-E) to a region corresponding to the second row of openings (112). Frames 16 to 20 capture a third sub-operation performed by the sealing machine (102) including positioning the first set of containers (104A-E) in a stationary position in the region corresponding to the second row of openings (112) for a designated duration during which the heat pistons (106A-E) heat the sealing regions (114K-O) and covers (105K-O) of the second set of containers (104K-O). Frames 21 to 30 capture a fourth sub-operation performed by the sealing machine (102) including moving the first set of containers (104A-E) from the stationary position, for example, to an associated assembly line or packaging conveyor (not shown in FIGS. 1-6) in the sealing machine (102).
[0057] Subsequently, at step (206), the frame selection system (122) identifies a first subset of frames in the first RGB video whose associated blur scores are greater than a designated blur score as blur-free frames. In one embodiment, the frame selection system (122) previously identifies the designated blur score using an open computer vision technique such as a fast Fourier transform based blur detection technique or a Laplacian blur detection technique. Further, the frame selection system (122) stores the designated blur score in an associated database (not shown in FIGS. 1-6) for identifying a first set of frames in the first RGB video that include blur-free images and a second set of frames in the first RGB video that include blurred images. For example, the frame selection system (122) identifies the frames 16 to 20 in the first RGB video whose associated blur scores are greater than an exemplary designated blur score of 7 as blur-free frames. As previously noted, in one embodiment, the frames whose associated blur scores are greater than the exemplary designated blur score of 7 indicate that those frames are blur-free frames captured when the first set of containers (104A-E) are positioned in the stationary condition. Alternatively, the frames whose associated blur scores are lesser than the exemplary designated blur score of 7 indicate that those frames include blurred images, such as those captured when the first set of containers (104A-E) are in motion.
[0058] At step (208), the frame selection system (122) determines unique identifiers (IDs) of the first subset of frames that are identified from the first RGB video. In one example implementation, the unique IDs of the first subset of frames correspond to chronological positions or sequence numbers of the first subset of frames in the first RGB video. For example, the unique IDs of the first subset of frames correspond to the numbers 16 to 20, which represent corresponding positions of the first subset of frames in the first RGB video.
[0059] At step (210), the frame selection system (122) identifies a second subset of frames that are blur-free frames in the second thermographic video based on the determined unique IDs of the first subset of frames. As previously noted, the first RGB video and the second thermographic video of the first set of containers (104A-E) is captured by the first imaging sensor (118) and the second imaging sensor (120) simultaneously at the same frame rate when the first set of containers (104A-E) is positioned in a stationary condition. Thus, each of the image frames in the first RGB video corresponds to a similar image frame in the simultaneously acquired second thermographic video. Accordingly, in the previously noted example where unique IDs of the first subset of frames were determined as 16 to 20, the frame selection system (122) identifies frames corresponding to 16th to 20th frame in the second thermographic video as the second subset of frames that are free from blur-related video errors.
[0060] Subsequently, at step (212), the frame selection system (122) selects any one frame from the second subset of frames and provides the selected frame as an input to the frame processing system (124). At step (214), the frame processing system (124) processes the selected frame for identifying if one or more of the first set of containers (104A-E) include a defective seal. At step (216), the defective container discarding system (128) automatically moves one or more of the first set of containers (104A-E) identified to include a defective seal from the sealing machine (102) to one or more defective container collection trays. Certain exemplary embodiments describing processing the selected frame for identifying if one or more of the first set of containers (104A-E) include defective seals, and subsequently discarding the containers including the defective seals are described in detail subsequently with reference to FIGS. 3-6.
[0061] In particular, FIGS. 3A-B illustrate a flowchart depicting an exemplary method (300) for processing the selected frame and for identifying if one or more of the first set of containers (104A-E) include defective seals by processing the selected frame using the seal inspection system (100). For simplicity, an embodiment of the present method (300) is described herein to identify if a particular container (104A) in the first set of containers (104A-E) includes a defective seal using the selected frame. However, it is to be understood that the method (300) can be similarly used to identify if other containers (104B-E) in the first set of containers (104A-E) include defective seals based on information determined by processing the selected frame.
[0062] At step (302), the frame processing system (124) receives the selected frame from the second subset of images as an input from the frame selection system (122). An exemplary selected frame (402) that is received as the input by the frame processing system (124) is depicted in FIG. 4A. The selected frame (402) includes, for example, an image of the sealing region (114A) including the cover (105A) of the container (104A).
[0063] At step (304), the frame processing system (124) receives a maximum temperature value and a minimum temperature value in the selected frame (402) as an input from the thermographic camera (120). In one embodiment, the thermographic camera (120) determines a corresponding maximum temperature value and a corresponding minimum temperature value in each frame in the second thermographic video. When the frame selection system (122) selects a particular frame from the second subset of frames, the frame selection system (122) provides a unique identifier, for example, a frame number associated with the selected frame as an input to the thermographic camera (120). Subsequently, the thermographic camera (120) provides the maximum temperature value and the minimum temperature value in the selected frame as the input to the frame processing system (124).
[0064] At step (306), the frame processing system (124) identifies a first pixel intensity value corresponding to the maximum temperature value and a second pixel intensity value corresponding to the minimum temperature value. At step (308), the frame processing system (124) interpolates a corresponding temperature value for each pixel in the selected frame (402) using a stored temperature-pixel intensity value correlation based on the maximum temperature value, the first pixel intensity value, the minimum temperature value, and the second pixel intensity value. In one example implementation, the frame processing system (124) is a SciPy library-based system that interpolates the corresponding temperature value for each pixel in the selected frame (402) using the maximum temperature value, the first pixel intensity value, the minimum temperature value, and the second pixel intensity value.
[0065] At step (310), the frame processing system (124) segregates a set of pixels in the selected frame (402) whose associated temperature values are greater than a designated temperature threshold as a hot region that corresponds to the sealing region (114A) of the container (104A). An example of the designated temperature threshold is 100° Celsius.
[0066] Subsequently, at step (312), the frame processing system (124) crops the hot region from the selected frame (402) and converts the cropped portion including the set of segregated pixels into a binary image frame. An exemplary cropped portion (404) that includes the set of segregated pixels indicative of the hot region is depicted in FIG. 4B. Further, the cropped portion (404) that is converted into an exemplary binary image frame (406) is depicted in FIG. 4C.
[0067] Alternatively, when the frame processing system (124) identifies that temperature values of all pixels in the selected frame (402) are lesser than the designated temperature threshold of 100° Celsius, the frame processing system (124) identifies that the container (104A) completely lacks a seal. For example, FIG. 4D depicts an exemplary selected frame (408) that includes an image of the sealing region (114A) of the container (104A). In one implementation, FIG. 4D may represent that the temperature values of all pixels in the selected frame (402) vary between 22.7° and 28.9° Celsius, the temperature values being lesser than the designated temperature threshold of 100° Celsius. When the sealing machine (102) does not adequately heat the sealing region (114A) and the cover (105A), the cover (105A) will not affix to the sealing region (114A) of the container (104A), and therefore, the container (104A) will remain open without any seal. Accordingly, in this example, the frame processing system (124) identifies that the container (104A) completely lacks a seal when temperature values of all pixels in the selected frame (408) are lesser than the designated temperature threshold of 100° Celsius.
[0068] Further, at step (314), the defect classifying system (126) that is pre-trained to identify one or more of the first set of containers (104A-E) including a defective seal classifies the binary image frame (406) corresponding to the hot region to identify if the container (104A) includes a defective seal. In one embodiment, the defect classifying system (126) is a machine learning system that employs one or more machine learning models such as one or more convolutional neural network models to identify if the container (104A) includes a defective seal such as a partial seal from the binary image frame (406).
[0069] To that end, the defect classifying system (126) is pre-trained with a first set of reference binary image frames of containers that include only partial seals and a second set of reference binary image frames of containers that include complete and proper seals. An exemplary first set of reference binary image frames (502A-F) and an exemplary second set of reference binary image frames (504A-F) that are used to pre-train the defect classifying system (126) are depicted in FIGS. 5A-F and FIGS. 5G-L, respectively.
[0070] During a training stage, the defect classifying system (126) learns patterns of partial or incomplete seals and non-uniform thickness seals from the first set of reference binary image frames (502A-F) and further learns patterns of complete and proper seals from the second set of reference binary image frames (504A-F). Post training, when the defect classifying system (126) receives the binary image frame (406) depicted in FIG. 4C as input from the frame processing system (124), the defect classifying system (126) identifies if the container (104A) includes a proper complete seal or a defective partial seal based on the different patterns of different types of seals learnt during the training stage.
[0071] For example, when the defect classifying system (126) receives the binary image frame (406) depicted in FIG. 4D as an input from the frame processing system (124), the defect classifying system (126) identifies the container (104A) to be a defective container and including only a partial seal based on patterns of partial seals learnt during the training stage. Alternatively, when the defect classifying system (126) receives another binary image frame that resembles one of the reference binary image frames (504A-F) as an input from the frame processing system (124), the defect classifying system (126) identifies the container (104A) to be a defect-free container with a proper seal based on patterns of complete seals learnt during the training stage.
[0072] At step (316), the defective container discarding system (128) determines a distance between a defective container picking unit (shown in FIG. 6) that is coupled to the sealing machine (102) and the container (104A) identified to include the defective seal, for example, based on a pre-calibration of the defective container discarding system (128). Subsequently, at step (318), the defective container discarding system (128) moves the defective container picking unit by the determined distance to move the container (104A) identified to include the defective seal from the sealing machine (102) to a defective container collection tray. An exemplary embodiment of the defective container discarding system (128) for use in moving the container (104A) identified to include the defective seal from the sealing machine (102) to a defective container collection tray is described in greater detail with reference to FIG. 6.
[0073] FIG. 6 illustrates a schematic diagram depicting an exemplary defective container discarding system (128) that moves one or more of the first set of containers (104A-E) identified to include a defective seal from the sealing machine (102) to one or more defective container collection trays (602A-D). In one embodiment, the defective container discarding system (128) includes a defective container picking unit controller (604), a set of motors (606A-D), a set of defective container picking units (608A-D), and a set of suction control valves (610A-D).
[0074] An exemplary operation of the defective container discarding system (128) to move the container (104A) identified to include a defective seal from the sealing machine (102) to the defective container collection tray (602A) is described in detail subsequently. However, it is to be understood that the defective container discarding system (128) can similarly move any other containers (104B-J) identified to include defective seals from the sealing machine (102) to one or more of the defective container collection trays (602A-D).
[0075] In one embodiment, the defective container picking unit (DCPU) controller (604) receives an input including a current position of the container (104A) in the sealing machine (102) from the thermographic camera (120). The thermographic camera (120) identifies the current location of the container (104A) including a particular row and a particular column in which the container (104A) is positioned in the sealing machine (102). For example, in the exemplary scenario depicted in FIG. 6, the thermographic camera (120) identifies that the container (104A) is positioned in a particular position (612) corresponding to a first row (614) and a first column (616) in the moveable tray (108). Subsequently, the thermographic camera (120) provides an input including the identified position (612) of the container (104A) in the sealing machine (102) to the DCPU controller (604).
[0076] In certain embodiments, the DCPU controller (604) is previously calibrated to store a number of motor revolutions required to move the defective container picking unit (608A) from an associated default position (618) to different positions corresponding to the first row (614) in the moveable tray (108) and from different positions in the first row (614) to the defective container collection tray (602A). In one embodiment, the DCPU controller (604) stores a number of motor revolutions required to move the defective container picking unit (608A) from the associated default position (618) in the sealing machine (102) to the identified position (612) corresponding to the first row (614) and the first column (616) in the moveable tray (108). The DCPU controller (604) controls the operation of the motor (606A) to rotate the motor (606A) in a first direction by only the stored number of motor revolutions. Rotating the motor in the first direction by only the stored number of motor revolutions causes the defective container picking unit (608A) to move and precisely position itself in the identified position (612) adjacent to the container (104A).
[0077] For simplicity, the defective container picking unit (608A) is subsequently described as a vacuum tube (608A) that sucks the container (104A) positioned on the moveable tray (108) to remove the container (104A) from the sealing machine (102). However, the defective container picking unit (608A) may also be a robotic arm that picks and removes the container (104A) from the sealing machine (102). Upon moving and precisely positioning the vacuum tube (608A) adjacent to the container (104A), the DCPU controller (604) opens up the suction control valve (610A). Additionally, the vacuum tube (608A) sucks the container (104A) positioned in the identified position (612) in the moveable tray (108) and removes the container (104A) from the sealing machine (102). Subsequently, the DCPU controller (604) controls the operation of the motor (606A) to rotate the motor (606A) in a second direction opposite to the first direction by a stored number of motor revolutions such that the vacuum tube (608A) carries the container (104A) and positions itself in the associated default position (618) in the sealing machine (102). The DCPU controller (604) then closes the suction control valve (610A) such that the vacuum tube (608A) drops the carried container (104A) into the defective container collection tray (602A).
[0078] For simplicity, each row in the sealing machine (102) is depicted to include a single vacuum tube (608A-D). However, it is to be understood that each row in the sealing machine (102) may include any number of vacuum tubes. For example, each row in the sealing machine (102) may include two or three vacuum tubes. In exemplary scenarios where a particular row in the sealing machine (102) includes more than one defective container, having multiple vacuum tubes in each row helps to remove all defective containers in that particular row quickly from the sealing machine (102).
[0079] The seal inspection system (100) described herein identifies one or more of the containers (104A-J) including a defective seal and automatically discards containers identified to include defective seals from the sealing machine (102). The present seal inspection system (100), thus, is capable of identifying containers including partial seals that include cuts or openings. The seal inspection system (100) is also capable of identifying a container having a leak-free seal, but whose width is not uniform throughout, and thus, needs to be discarded as a defective container as such a non-uninform thickness and leak-free seal has a good chance of being failure prone at a later point of time. Specifically, the non-uniform thickness and leak-free seal is prone to fail and tear apart at a portion where the sealing is thin, and thereby exposes the content or item stored within a container and spoils the quality of the content or item. In other words, the seal inspection system (100) is capable of identifying the container including the leak-free seal whose width is thinner at one associated portion and is thicker at another associated portion. To that end, the frame processing system (124) in the seal inspection system (100) identifies a hot region corresponding to a sealing region of the container after application of heat on the sealing region and/or cover of the container by a heat piston (106A) of the sealing machine (102). In certain embodiments, the frame processing system (124) uses the method (300) described previously with reference to FIGS. 3A-B for identifying the hot region corresponding to the sealing region of the container.
[0080] In one embodiment, the hot region thus identified by the frame processing system (124) resembles a shape of two concentric circles similar to a binary image of an exemplary seal depicted in FIG. 5G. The defect classifying system (126) then verifies if a thickness of a region between the two concentric circles is even throughout the entire region circumferentially. For example, the defect classifying system (126) identifies that the thickness of the region between the two concentric circles is 6 mm and is even throughout the entire region circumferentially. In this example, the defect classifying system (126) identifies that the container to include a defect-free seal. However, in another example, the defect classifying system (126) identifies that the thickness of one portion of the region is 6 mm and the thickness of another portion of the region is only 3 mm. In this example, the defect classifying system (126) identifies that the container includes a leak-free seal but associated width is not uniform throughout. Subsequently, the defective container discarding system (128) discards the container having the leak-free seal whose width is not uniform throughout.
[0081] Additionally, unlike conventional seal inspection systems, embodiments of the seal inspection system (100) presented herein allow for simultaneous sealing and seal inspection-related imaging operations, thereby ensuring greater productivity and capacity for container packaging facilities. In particular, the seal inspection system (100) includes thermographic camera (120) that continuously captures images of the first set of containers (104A-E) in a stationary position for better clarity, while the sealing machine continues to seal the second set of containers (104K-O), which improves productivity and a number of containers sealed over a period of time.
[0082] Further, unlike conventional seal inspection systems, the thermographic camera (120) used in the seal inspection system (100) does not need frequent re-calibration as the thermographic camera (120) continuously captures images of the containers (104A-O). Further, the seal inspection system (100) identifies blur-free frames in a video that is continuously captured by the thermographic camera (120) and uses only those identified blur-free frames for inspecting the seal quality and integrity, thus improving accuracy with which containers including defective seals are detected.
[0083] In certain embodiments, the seal inspection system (100) also notifies a concerned person of one or more of the heat pistons (106A-E) that are malfunctioning, thereby assisting the concerned person to carry out preventive maintenance of the sealing machine (102) in a timely manner. For example, the defect classifying system (126) identifies that every single container (104A) that is heat sealed by the heat piston (106A) includes a defective seal. In this example, the seal inspection system (100) identifies that the defective seal may be occurring due to a change in an orientation of the heat piston (106A) or due to accumulation of impurities on the heat piston (106A). Accordingly, the seal inspection system (100) may present one or more of an audio and a video notification on the alerting system (130) including a display device and/or an audio-visual system to alert the concerned person of malfunctioning of the heat piston (106A) to carry out necessary corrective actions.
[0084] 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.
[0085] 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.

LIST OF NUMERAL REFERENCES:

100 Seal inspection system
102 Sealing machine
104A-O Containers
105A-E, 105K-O Container covers
106A-E Heat pistons
108 Moveable tray
110, 112 First and second row of openings
114A-J, 114K-O Sealing regions of containers
116 Direction of movement of tray
118 First imaging sensor
120 Second imaging sensor
122 Frame selection system
124 Frame processing system
126 Defect classifying system
128 Defective container discarding system
130 Alerting system
200-216 Steps of a method for identifying if containers include defective seals
300-318 Steps of a method for processing a frame selected from a video to identify if a container includes a defective seal
402 Exemplary selected frame
404 Cropped image
406 Binary image frame
502A-F Training images of containers including partial seals
504A-F Training images of containers including complete seals
602A-D Defective container collection trays
604 DCPU Controller
606A-D Motors
608A-D Defective container picking units
610A-D Suction control valves
612 Position of defective container
614 First row of moveable tray
616 First column of moveable tray
618 Default position of a defective container picking unit
,CLAIMS:We claim:

1. A seal inspection system (100), comprising:
a first imaging sensor (118) having a first type that captures a first video of a first container (104A) and a second imaging sensor (120) having a second type different from the first type that captures a second video of the first container (104A) while a sealing operation is being performed by a sealing machine (102) on a second container (104K), wherein the first video and the second video are captured by the first imaging sensor (118) and the second imaging sensor (120) simultaneously and at the same frame rate when the first container (104A) is positioned in a stationary condition;
a frame selection system (122), a frame processing system (124), and a defect classification system (126) operatively coupled to the first imaging sensor (118) and the second imaging sensor (120);
wherein the frame selection system (122) identifies a first subset of frames in the first video whose associated blur scores are greater than a designated blur score, and identifies a second subset of frames in the second video corresponding to the first subset of frames;
wherein the frame processing system (124) processes a selected frame (402) from the second subset of frames to identify a hot region that corresponds to a sealing region (114A) of the container (104A);
wherein the defect classification system (126) classifies the hot region based on different patterns of seals learnt during a training stage of the defect classifying system (126) to identify if the first container (104A) comprises a defective seal; and
an alerting system (130) communicatively coupled to the defect classification system (126) and adapted to generate an alert upon identifying that the first container (104A) comprises a defective seal.

2. The seal inspection system (100) as claimed in claim 1, wherein the first imaging sensor (118) corresponds to a red-green-blue camera, and wherein the second imaging sensor (120) corresponds to a thermographic camera.

3. The seal inspection system (100) as claimed in claim 1, wherein the first container (104A) corresponds to a sealable container for storing one or more of a food item, beverage, coffee powder, tea powder, a chemical product, an industrial product, an aerosol, and a medicinal product.

4. The seal inspection system (100) as claimed in claim 1, wherein the seal inspection system (100) comprises a defective container discarding system (128) that automatically removes the first container (104A) identified to comprise the defective seal from the sealing machine (102).

5. The seal inspection system (100) as claimed in claim 4, wherein the defective container discarding system (128) comprises a defective container picking unit controller (604), a set of motors (606A-D), a set of defective container picking units (608A-D), and a defective container collection tray (602A), wherein the set of defective container picking units (608A-D) comprises one or more of a robotic arm and a vacuum tube (608A-D) adapted to automatically move the first container (104A) identified to comprise the defective seal from the sealing machine (102) to the defective container collection tray (602A).

6. The seal inspection system (100) as claimed in claim 1, wherein the alerting system (130) comprises one or more of a display device and an audio-visual system, wherein the alerting system (130) presents one or more of an audio and a video notification to alert an operator of the sealing machine (102) regarding a count of defective containers (104A-E), a position of one or more of the defective containers (104A-E), a type of the identified defect, and an issue with the sealing machine (102).

7. A method for inspecting a seal of a container, comprising:
capturing a first video of a first container (104A) by a first imaging sensor (118) having a first type and a second video of the first container (104A) by a second imaging sensor (120) having a second type different from the first type while a sealing operation is being performed by a sealing machine (102) on a second container (104K), wherein the first video and the second video are captured by the first imaging sensor (118) and the second imaging sensor (120) simultaneously at the same frame rate when the first container (104A) is positioned in a stationary condition;
determining a corresponding blur score for each frame in the first video and identifying a first subset of frames in the first video whose associated blur scores are greater than a designated blur score;
identifying a second subset of frames in the second video corresponding to the first subset of frames and selecting a frame (402) from the identified second subset of frames;
identifying a hot region in the selected frame (402) that corresponds to a sealing region (114A) of the container (104A);
classifying the hot region based on different patterns of seals learnt during a training stage of an associated defect classifying system (126) to identify if the first container (104A) comprises a defective seal; and
generating an alert by an alerting system (130) upon identifying that the first container (104A) comprises a defective seal.

8. The method as claimed in claim 7, wherein identifying a second subset of frames in the second video comprises:
determining unique identifiers of the first subset of frames, wherein the determined unique identifiers correspond to sequence numbers of the first subset of frames in the first video; and
selecting frames from the second video that correspond to the sequence numbers of the first subset of frames to identify the second subset of frames.

9. The method as claimed in claim 7, wherein identifying the hot region in the selected frame (402) comprises:
interpolating a corresponding temperature value for each pixel in the selected frame (402) using a stored temperature-pixel intensity value correlation based on a maximum temperature value in the selected frame (402), a minimum temperature value in the selected frame (402), a first pixel intensity threshold corresponding to the maximum temperature value, and a second pixel intensity corresponding to the minimum temperature value;
segregating a set of pixels in the selected frame (402) whose associated temperature values are greater than a designated temperature threshold as the hot region;
cropping the set of segregated pixels indicative of the hot region from the selected frame (402); and
converting a cropped portion (404) comprising the set of segregated pixels into a binary image frame (406) and sharing the binary image frame (406) with the defect classification system (126).

10. The method as claimed in claim 7, wherein the method comprises training the defect classification system (126) to learn one or more patterns indicative of one or more types of defective seals, wherein the one or more types of defective seals comprises an incomplete seal and a seal comprising non-uniform thickness, wherein the defect classification system (126) learns one or more patterns indicative of incomplete seals using a first set of reference binary image frames (502A-F), and learns one or more patterns indicative of one or more of complete seals using a second set of reference binary image frames (504A-F).

11. The method as claimed in claim 7, wherein the method comprises automatically moving the first container (104A) identified to comprise the defective seal from the sealing machine (102) to a defective container collection tray (602A) by a defective container discarding system (128).

12. The method as claimed in claim 11, wherein the method comprises:
pre-calibrating the defective container discarding system (128) to store specific numbers of motor revolutions required to move an associated defective container picking unit (608A) from a default position (618) to different positions in the moveable tray (108) in the sealing machine (102) and from the different positions in the moveable tray (108) to the defective container collection tray (602A);
rotating the motor (606A) in a first direction by a stored specific number of motor revolutions to position the defective container picking unit (608A) in an identified position (612) adjacent to the first container (104A) identified to comprise a defective seal;
removing the first container (104A) from the sealing machine (102) by the associated defective container picking unit (608A); and
rotating the motor (606A) in a second direction opposite to the first direction by a stored specific number of motor revolutions such that the associated defective container picking unit (608A) moves the first container (104A) into the defective container collection tray (602A).