Claims:1. A system (100) for automatic target recognition, said system comprising:
a server (102) that detects arrival of input data from earth station (106);
a plurality of automatic target recognition (ATR) clients (104) coupled to said server, said server configured to:
receive the input data pertaining to images from the earth station;
process the input data to analyse area of interest for each geographical location and discard the non-useful area; and
generate the processed reports and transmit to respective ATR client automatically based on area of responsibility of the corresponding ATR clients associated with area of interest, wherein the corresponding ATR clients transmits generated feedback to the server for updating target signature.
2. The system as claimed in claim 1, wherein said plurality of ATR clients (104) generate feedback using an online feedback loop, wherein the target signature in library is updated for false positive, miss classified targets, and addition of any new signature.
3. The system as claimed in claim 1, wherein said server (102) calculates the reliability score for target signatures in the library based on the feedback received from the respective ATR clients.
4. The system as claimed in claim 3, wherein the reliability score is being calculated based on the number of times a particular target signature is responsible for generation of correct target detections, false target detections and miss classifications.
5. The system as claimed in claim 1, wherein the feedback received from the respective ATR clients (104) ensure reduction of false alarms by retaining useful target signatures in the library.
6. The system as claimed in claim 1, wherein the server performs hybridized adaptive threshold-based template matching and feature matching approach for target recognition.
7. The system as claimed in claim 1, wherein the adaptive thresholds for various confidence levels is used for processing the input data at the server.
8. The system as claimed in claim 7, wherein the template matching provides true positive targets when the confidence level of detection is higher than adaptive higher threshold, wherein to control the false alarm rate, the template matching is applied for the set of results having confidence level below the adaptive higher threshold and above the adaptive lower threshold using unknown class templates.
9. The system as claimed in claim 8, wherein the false targets are discarded and the remaining targets are subjected to feature matching for discarding missed false targets.
10. A method (300) for automatic target recognition, said method comprising:
receiving (302), at a server, input data pertaining to images from earth station, the server detects arrival of input data from the earth station;
processing (304), at the server, the input data to analyse area of interest for each geographical location and discard the non-useful area; and
generating (306), at the server, the processed reports and transmit to respective plurality of automatic target recognition (ATR) clients automatically based on area of responsibility of the corresponding ATR clients associated with area of interest, the plurality of ATR clients coupled to the server wherein the corresponding ATR clients transmits generated feedback to the server for updating target signature.
, Description:TECHNICAL FIELD
[0001] The present disclosure relates, in general, to automatic image analysis, and more specifically, relates to a system and method for automatic target recognition and classification from satellite imagery.

BACKGROUND
[0002] Satellite imagery provides a wide range of information and they are of particular use for defence applications for recognition and classification of targets and activities, extraction of their geo-locations, and intelligence gathering. In most of the image processing and analysis systems that exist today, the images received from the sensors are stored in an image database and from there the images are fetched into image processing and analysis systems/machines through various advanced searching methods and images are processed and analysed and the results, as well as reports, are provided to the users.
[0003] There has been remarkable progress and technological advances in imaging sensors, image database design and development, and with the same pace in image processing and analysis algorithms in recent years. However, in almost all the cases, the image searching and selection, selection of processing and analysis algorithms, and processing and result generation is carried out manually by the operators and the image processing and analysis systems or machines operate in a standalone manner. There is hardly any automatic image processing and analysis workflow that covers the entire sensor to user loop, wherein the images are processed and analysed online at the back end as and when they are captured and received from the sensor site and processing and analysis results are automatically provided to the respective users in real-time in the form of processed outputs, information and alerts.
[0004] The type of targets detected by the system are configured into the system and the whole algorithm is designed to develop a solution for detecting a particular type of target. It was seen that when dealing with a variety of targets, settling upon a unique quantifiable confidence factor from the system point of view is a wrong assumption. Some targets were adjudged correctly at even lower confidence levels of the system while at the same time some targets required a higher confidence level.
[0005] An existing system in the field of satellite imagery includes a direct broadcasting imaging satellite system and method for providing real-time, continuous monitoring of earth from geostationary earth orbit and broadcasting it to end-users. However, this method and system for image collection and transfer, does not relate to image processing, analysis and transfer to required end-users. Another existing system includes an image capturing and processing system that supports multitier modular software, and plug-in extendable, architecture, however, this system does not relate to an architecture wherein, there are multiple sources and different types of data. The system does not relate to a distributed multitier architecture where the end-user is located at different geo-locations. Another existing method and apparatus relate to real-time target recognition within a multispectral image with a sparsely driven target recognition algorithm utilizing a set of parameters tuned with a deep neural network, however, it does not mention various kinds of imagery, automatic image correction and thereafter automatic processing and distribution of tailored results to end-users. Yet another existing method include target recognition and identification in the OCR field. However, the method does not mention varying the confidence factor itself for different categories of targets.
[0006] A few existing literature items have been studied by the researchers, however, an extensive mechanism was not found which relates to the architectural requirements for processing and analysis of images in detail, right from the processing of the assimilated data from various sources and dissemination of results as a single autonomous system. The existing technologies known in the art relates to the capturing of the data part or the processing of the data part, however, there is no such application or framework, which talks about a complete architecture or framework requirement. In addition, processing the entire input image leads to slower processing and less reliable results. The result is displayed on the screen of the end-user. However, these results might not always be useful for the end-user and may lead to the cluttering of results. The existing technologies also discuss the online feedback loop. The feedback does not affect the results instantly, as the system needs to be free from the current processing task.
[0007] In the existing methods, where the target recognition is done perform the task by using a generic confidence factor for deciding upon the presence or absence of target. This is mostly due to the fact that the shape and size of the targets are uniform or it is made uniform by scaling the target image to a predefined size. This is also mostly done in the recent artificial intelligence (AI) based systems that are being developed. The existing methods do not provide varying confidence factors for different categories of targets present in the same image.
[0008] Therefore, it is desired to develop a unique automatic workflow based real time image processing and analysis system using adaptive threshold technique for detection of different kind of targets.

OBJECTS OF THE PRESENT DISCLOSURE
[0009] An object of the present disclosure relates in general, to automatic image analysis, and more specifically, relates to a system and method for automatic target recognition and classification from satellite imagery.
[0010] Another object of the present disclosure is to provide an integrated autonomous system with a complete workflow starting from automatic scanning for detection of new inputs, then analyzing and eventually displaying useful results.
[0011] Another object of the present disclosure is to provide a system that provides the arrival of data that triggers the automatic processing and analysis operations, where the analysis results are spontaneously distributed.
[0012] Another object of the present disclosure is to provide a system that provides optimized workflow, which ensures fast processing and analysis of data that ensures timely reporting of important results to the designated clients.
[0013] Another object of the present disclosure is to provide a system that provides adaptive thresholding that helps in improving the results.
[0014] Another object of the present disclosure is to provide a system that provides an online feedback mechanism that improves the results on the fly.
[0015] Another object of the present disclosure is to provide a system that provides automatic nature of features, helps the human operator to speed up the task.
[0016] Another object of the present disclosure is to provide a system that enables in choosing the most beneficial target signatures for performing ATR while working on a new image and thus improving results.
[0017] Another object of the present disclosure is to provide a system that enables the automatic area of responsibility (AOR) based image processing for target detection for the reduction in the size of the image to be processed at the server.
[0018] Yet another object of the present disclosure is to provide a generic architecture framework for image processing and image analysis based on libraries and contextual databases within command, control, communications, computers, and intelligence (C4I) systems/ environment.

SUMMARY
[0019] The present disclosure relates, in general, to automatic image analysis, and more specifically, relates to a system and method for automatic target recognition and classification from satellite imagery.
[0020] In an aspect, the present disclosure relates to a system for automatic target recognition, the system comprising a server that detects arrival of input data pertaining to image from earth station, a plurality of ATR clients coupled to the server, the server configured to receive the input data from the earth station, process the input data to analyse area of interest for each geographical location and discard the non-useful area, and generate the processed reports and transmit to respective ATR client automatically based on area of responsibility of the corresponding ATR clients associated with area of interest, wherein the corresponding ATR clients transmits generated feedback to the server for updating target signature.
[0021] According to an embodiment, the plurality of ATR clients generate feedback using an online feedback loop, wherein the target signature in library is updated for false positive, miss-classified targets, and addition of any new signature.
[0022] According to an embodiment, the server calculates the reliability score for target signatures in the library based on the feedback received from the respective ATR clients.
[0023] According to an embodiment, the reliability score is being calculated based on the number of times a particular target signature is responsible for generation of correct target detections, false target detections and miss classifications.
[0024] According to an embodiment, the feedback received from the respective ATR clients ensure reduction of false alarms by retaining useful target signatures in the library.
[0025] According to an embodiment, the server can perform hybridized adaptive threshold-based template matching and feature matching approach for target recognition.
[0026] According to an embodiment, the adaptive thresholds for various confidence levels is used for processing the input data at the server.
[0027] According to an embodiment, the template matching provides true positive targets when the confidence level of detection is higher than adaptive higher threshold, wherein to control the false alarm rate, the template matching is applied for the set of results having confidence level below the adaptive higher threshold and above adaptive lower threshold using unknown class templates.
[0028] According to an embodiment, the false targets are discarded and the remaining targets are subjected to feature matching for discarding missed false targets.
[0029] In an aspect, the present disclosure relates to a method for automatic target recognition, the method includes receiving, at a server, input data pertaining to images from earth station, the server detects arrival of input data from the earth station, processing, at the server, the input data to analyse area of interest for each geographical location and discard the non-useful area and generating, at the server, the processed reports and transmit to respective plurality of ATR clients automatically based on area of responsibility of the corresponding ATR clients associated with area of interest, the plurality of ATR clients coupled to the server wherein the corresponding ATR clients transmits generated feedback to the server for updating target signature.
[0030] Various objects, features, aspects, and advantages of the inventive subject matter will become more apparent from the following detailed description of preferred embodiments, along with the accompanying drawing figures in which like numerals represent like components.

BRIEF DESCRIPTION OF THE DRAWINGS
[0031] The following drawings form part of the present specification and are included to further illustrate aspects of the present disclosure. The disclosure may be better understood by reference to the drawings in combination with the detailed description of the specific embodiments presented herein.
[0032] FIG. 1A illustrates an exemplary system for real-time image processing and analysis, in accordance with an embodiment of the present disclosure.
[0033] FIG. 1B illustrates a flow diagram depicting sequence of operations performed in the proposed autonomous framework, in accordance with an embodiment of the present disclosure.
[0034] FIG. 2A illustrates a block diagram of the hybridized ATR solution, in accordance with an embodiment of the present disclosure.
[0035] FIG. 2B illustrates a flow chart for hybrid ATR, in accordance with an embodiment of the present disclosure.
[0036] FIG. 3 illustrates an exemplary flow chart of a method for automatic target recognition, in accordance with an embodiment of the present disclosure

DETAILED DESCRIPTION
[0037] The following is a detailed description of embodiments of the disclosure depicted in the accompanying drawings. The embodiments are in such detail as to clearly communicate the disclosure. If the specification states a component or feature “may”, “can”, “could”, or “might” be included or have a characteristic, that particular component or feature is not required to be included or have the characteristic.
[0038] As used in the description herein and throughout the claims that follow, the meaning of “a,” “an,” and “the” includes plural reference unless the context clearly dictates otherwise. Also, as used in the description herein, the meaning of “in” includes “in” and “on” unless the context clearly dictates otherwise.
[0039] The present disclosure relates, in general, to automatic image analysis, and more specifically, relates to a system and method for automatic target recognition and classification from satellite imagery. The present disclosure relates to the design and development of a satellite imagery analysis system based on an automatic image analysis workflow using a hybridized method for automatic target recognition and classification. In an embodiment, the system and method of the present disclosure enable to overcome the limitation of the prior art by enabling unique automatic workflow based real-time image processing and analysis system using adaptive threshold technique for detection of different kinds of targets. In this approach, the accuracy of the system achieved is judged as the type of targets identified correctly. The system identifies the type of targets with the confidence factor, which is configured in the system and is used as a unique defining factor for deciding the presence or absence of target.
[0040] The system and method of the present disclosure enable imagery, automatic image correction and thereafter automatic processing and distribution of tailored results to end-users and enable varying the confidence factors for different categories of targets. The threshold or confidence factor for different targets were made adaptable while looking for different categories of targets in the same satellite image.
[0041] The present disclosure relates to an automatic system for hybridized template and feature-based automatic target recognition using adaptive thresholding that explains the process of detection of the arrival of the new image at a central location, image processing and analysis, and distribution of image analysis results to the designated clients. The system and method meet the various requirement of automatic target recognition and classification. The automatic target recognition algorithm has been fine-tuned by hybridizing template and feature matching approaches and using adaptive thresholds for various confidence levels for target matching. The processed outputs from the target recognition method shall provide a comprehensive situational awareness input to the various operational and planning systems. The description of terms and features related to the present disclosure shall be clear from the embodiments that are illustrated and described; however, the invention is not limited to these embodiments only. Numerous modifications, changes, variations, substitutions, and equivalents of the embodiments are possible within the scope of the present disclosure. Additionally, the invention can include other embodiments that are within the scope of the claims but are not described in detail with respect to the following description.
[0042] FIG. 1A illustrates an exemplary system for real-time image processing and analysis, in accordance with an embodiment of the present disclosure.
[0043] Referring to FIG. 1A, system 100 configured for real-time image processing and analysis for target detection from satellite imagery. The system 100 can include server 102, automatic target recognition (ATR) clients 104-1 to 104-N (which are collectively referred to as ATR clients 104 and individually referred to as ATR client 104, hereinafter) and earth station 106. FIG. 1A depicts the availability of input data from earth station 106 to server 102 and ATR clients 104. The framework presented in the disclosure relates to a complete architecture describing the required entities and hence addressing the problem of creation of a complete situation picture in front of the distant end-user through automatic target extraction.
[0044] The input data from earth station 106 is received by server 102 and one or more ATR clients 104 (also referred to as ATR clients, 104 herein). The system 100 depicts the flow of analysed data from server 102 to ATR clients 104 and feedback from ATR clients 104 to server 102. The client-side application provides capabilities to the ATR analysts to execute various functionalities on the processed outputs including the customized report generation. The present disclosure includes advanced methods of ATR and classification from satellite panchromatic, multispectral and hyperspectral imagery. The system 100 focuses on the exploitation of an exhaustive template library, modelling of the ‘unknown’ class and hybridizing the template and feature-based approaches and adaptive thresholding for confidence levels, to provide significant improvement in target classification and extraction results as well as a higher degree of accuracy of ATR.
[0045] The proposed system 100 presents a detailed architecture for analysing the captured input data from various types of sensors, which include images from ground sensors, aircraft, unmanned aerial vehicles (UAVs) and satellites. The data captured is archived and is also available for analysis over the dedicated secure network. The data availability is time/requirement dependent. Thus, there is a need for procedures, which help in discovering the presence of data. Once the data or images are discovered, the same can be processed by server 102 housing the ATR engine. There are some specialized operations such as ortho-rectification, image registration, target detection, target extraction along with various pre-processing functions in the ATR engine. Also, a geo-database is present along with a standard database. The database is used for storing various knowledge base libraries, feature libraries for various kinds of targets.
[0046] As shown in FIG. 1A, server 102 can detect the arrival of input data from earth station 106. The continuous scanning of the shared drive, where input data arrives from the earth station 106 is being performed, for the detection of the arrival of new input data. The server 102 can process the input data. The detection of the arrival of new data, automatically triggers the processing and analysis of new input data, independent of the operator invocation. The server 102 can analyse the area of interest (AOI) for each geographical location in the automatic processing of input data, due to which the non-useful area has been excluded from getting processed and the false alarms have been reduced. In an exemplary embodiment, automatic target recognition may be improved by the incorporation of the concept of area of interest for each geographical location of which images are captured for the purpose of analysis and surveillance. This has also helped in optimizing the generation of results and results that are more reliable with less false detections due to the reduced area of operation.
[0047] The server 102 can generate processed reports and transmit to respective ATR clients 104 automatically based on an area of responsibility of the corresponding ATR clients 104 associated with area of interest. The ATR analysts are responsible for the analysis of specific geographical regions. Hence, only the necessary information is sent to respective ATR clients 104 from the centralized ATR engine. The incorporation of the area of responsibility has further addressed the problem of cluttering unnecessary reports at respective ATR clients 104. The automatic area of responsibility based image processing for target detection is performed for the reduction in the size of the image to be processed.
[0048] The ATR clients 104 generate feedback using an online feedback loop and transmit to the server 102 for updating the target signature in the library for false-positive and miss-classified targets, and caters to the addition of any new signature to the existing signature library. The online feedback loop is designed for the real-time consumption of the feedback as the target library has been stored in the form of target chips and features of the target chip instead of computing a model for extraction of targets, which requires fine-tuning of parameters and is time-consuming to generate. The processing algorithm is also fine-tuned to work on this kind of library.
[0049] The server 102 can calculate the reliability score for target signatures in the library based on the feedback received from respective ATR clients 104, which has helped in further reduction of false alarms by keeping the most useful target signatures in the library. This has helped in the automatic cleaning of the library. The score is being calculated based on the number of times a particular target signature is responsible for the generation of correct target detections, false target detections or miss-classifications. The use of feedback has been enhanced in the present disclosure and is being used not only by the processing algorithm but also for calculating the ranking of target chips, which act as a reliability score of target chips. The automatic ranking mechanism for ranking of target library signatures based on feedback from clients distributed geographically. This helps in choosing the most appropriate target signatures for performing ATR while working on a new image and thus improving results. The target chip is dropped from the library as soon as the reliability score of a particular target chip drops below a pre-defined reliability score. This helps in auto cleaning of the library, which otherwise would have grown enormously and led to the generation of less reliable results. Also, the time needed for the detection of the target in the image would have grown unchecked.
[0050] The hybridized adaptive threshold-based template matching and feature matching approach is used for target recognition, wherein the adaptive thresholds for various confidence levels have been used for processing the images at server 102. The template matching provides true positive targets when the confidence level of detection is higher than the adaptive higher threshold, wherein to control the false alarm rate, the template matching is applied for the set of results having confidence level below the adaptive higher threshold and above the adaptive lower threshold using only unknown class templates. Some false targets are discarded at this stage and the remaining targets are again subjected to feature-based ATR, for discarding missed false targets. The target list from feature-based ATR and template-based ATR are combined to get a final list of true positive targets. The target detection accuracy is dependent on the library which is used as a starting point for any such system. The better the library the sooner the results of such a system become useful for the end-user eventually reducing manual work.
[0051] The embodiments of the present disclosure described above provide several advantages. The present disclosure provides an integrated autonomous system 100 with a complete workflow starting from automatic scanning for detection of new inputs, then analyzing and eventually displaying useful results. The system 100 enables the arrival of data that triggers the automatic processing and analysis operations, where the analysis results are spontaneously distributed. The system 100 provides optimized workflow, which ensures fast processing and analysis of data that ensures timely reporting of important results to the designated clients. The system provides adaptive thresholding that helps in improving the results, provides the online feedback mechanism that improves the results on the fly, provides automatic nature of features and helps the human operator to speed up the task. The proposed system 100 enables in choosing the most beneficial target signatures for performing ATR while working on a new image and thus improving results. The system 100 enables the automatic area of responsibility based image processing for target detection for the reduction in the size of the image to be processed at the server and provides a generic architecture framework for image processing and image analysis based on libraries and contextual database within command, control, communications, computers, and intelligence (C4I) systems/ environment.
[0052] FIG. 1B illustrates a flow diagram depicting sequence of operations performed in the proposed autonomous framework, in accordance with an embodiment of the present disclosure.
[0053] Referring to FIG. 1B, the complete workflow of the function performed at the ATR engine is presented in FIG 1B. Once the results are computed and transmitted to respective clients, the same is analysed by client 104 and consumed for taking important decisions related to the security of the surveillance area. The feedback pertaining to the computed results is an important step towards the success of any such system. Thus, it has been given special importance and has been recorded for updating the wrong targets reported and for defining new targets. The feedback has further been utilized for improving the overall quality of the target library.
[0054] The special reliability score has been calculated for each target signature captured in the library. The reliability score is calculated based on the number of times a target signature is responsible for true positive and false-positive alerts. This reliability score helps in reducing the false targets reported by the system 100. This reliability score also helps in automatic cleaning of the target library by simply dropping a target signature once the reliability score drops below a particular pre-defined threshold. This also helps in auto removal of targets added by mistake by the clients. Apart from the above flow, AOI selection, defining clients responsible for AOI(s) and defining thresholds for resolution scaling of target templates with the respect target image, can also be configured as per need.
[0055] At block 108, scanning the data for input, at block 110 only the metadata files are read initially to check for AOI. At block 112, the image is further read only if the AOI lies in the image. The image is processed only for the portion within AOI, rather than the whole input image. Also, dynamic multithreading is used in processing the image. The number of threads is dynamically created based on the target library. This helps in optimized results in terms of both time and accuracy. At block 114, these results are forwarded to clients as per the responsibility of clients for particular AOI(s). At block 116, the ATR clients 104 generate feedback and sent to server 102 for updating the target signature in the library.
[0056] FIG. 2A illustrates a block diagram 200 of the hybridized ATR solution, in accordance with an embodiment of the present disclosure. FIG. 2A depicts the hybridization of template and feature based approaches.
[0057] Referring to FIG. 2A, a specialized hybridized algorithm for template and feature matching based on adaptive thresholding has been used for ATR. The block diagram depicts the ATR solution at the abstract level. The accuracy of template matching for high confidence levels is high, but for lower confidence levels, the probability of false detections increases. Hence, template-based ATR results are refined by the inclusion of an unknown class in template and feature libraries, in addition to the target library comprising of classes for different types of targets.
[0058] The modelling of the unknown class, where the target resembles the background imagery to a large extent is based on automatic analysis of imagery. For the creation of an unknown class of library, the template matching technique is applied on the target image using template library created initially without the unknown class. The target library is defined by cutting target chips from the images and extracting the features of target chips. The result is analysed visually and false-positive classifications are grouped through a feedback mechanism in a new class of library known as ‘unknown’. Also, the same feedback loop is used to capture the miss-classifications and false negatives. The ATR engine supports updating of the library as an iterative process, where the scope to improve the library is left open-ended, which helps in improving the accuracy of ATR at any stage. The accuracy of the hybridized algorithm mentioned above is further fine-tuned using adaptive thresholding at various stages of the algorithm. The adaptive thresholds have been chosen based on the type of targets being detected.
[0059] FIG. 2B illustrates a flow chart for hybrid ATR, in accordance with an embodiment of the present disclosure. The flow chart depicts the hybridized template and feature based matching methods using adaptive thresholds. At block 202, in the template matching technique, all the target classes except the unknown class are used for analysing the image. It is normally observed in template matching that a high confidence of detection C1 always relates to a true positive target i.e., confidence of detection C(i) is higher than adaptive higher threshold C1. In other words, the false alarms rate is inversely proportional to the value of confidence of detection C (i). At block 204, adaptive higher threshold C1 varies according to the target type. Thus, in order to control the false alarm rate, template based matching is applied for the set of results having confidence level below the adaptive higher threshold C1 and above the adaptive lower threshold C2 for ATR using only unknown class templates. At block 206, if the confidence achieved in matching with unknown class is higher than previously found confidence level, then such targets are removed from the result set.
[0060] At block 208, the remaining targets are again subjected to feature-based ATR. At block 210, if the distance value of the unknown class is found to be lesser than other classes, the target is deleted from the result set, where d1 distance of target from library and d2 distance of target from unknown class. Hence, the false alarm rate is kept under control at lower confidence levels of the template matching approach. Further, the results with an even lower confidence level C2 in template matching are discarded straight forward without any analysis. In other words, they are assumed to be false alarms. At block 212, list from template-based ATR and list from feature-based ATR can be combined to generate the target results.
[0061] FIG. 3 illustrates an exemplary flow chart of a method for automatic target recognition, in accordance with an embodiment of the present disclosure.
[0062] Referring to FIG. 3, the method 300 includes at block 302, the server can receive input data pertaining to images from earth station, the server detects arrival of input data from the earth station. At block 304, the server can process the input data to analyse area of interest for each geographical location and discard the non-useful area.
[0063] At block 306, the server can generate the processed reports and transmit to respective plurality of ATR clients automatically based on area of responsibility of the corresponding ATR clients associated with area of interest, the plurality of ATR clients coupled to the server, where the corresponding ATR clients transmits generated feedback to the server for updating target signature.
[0064] It will be apparent to those skilled in the art that the system 100 of the disclosure may be provided using some or all of the mentioned features and components without departing from the scope of the present disclosure. While various embodiments of the present disclosure have been illustrated and described herein, it will be clear that the disclosure is not limited to these embodiments only. Numerous modifications, changes, variations, substitutions, and equivalents will be apparent to those skilled in the art, without departing from the spirit and scope of the disclosure, as described in the claims.

ADVANTAGES OF THE PRESENT DISCLOSURE
[0065] The present disclosure provides an integrated autonomous system with a complete workflow starting from automatic scanning for detection of new inputs, then analyzing and eventually displaying useful results.
[0066] The present disclosure provides a system that enables the arrival of data triggers the automatic processing and analysis operations, where the analysis results are spontaneously distributed.
[0067] The present disclosure provides a system that provides optimized workflow which ensures fast processing and analysis of data that ensures timely reporting of important results to the designated clients.
[0068] The present disclosure provides a system that provides adaptive thresholding that helps in improving the results.
[0069] The present disclosure provides a system that provides an online feedback mechanism that improves the results on the fly.
[0070] The present disclosure provides a system that provides automatic nature of features, helps the human operator to speed up the task.
[0071] The present disclosure provides a system that enables in choosing the most beneficial target signatures for performing ATR while working on a new image and thus improving results.
[0072] The present disclosure provides a system that enables the automatic area of responsibility (AOR) based image processing for target detection for the reduction in the size of the image to be processed at the server.
[0073] The present disclosure provides a generic architecture framework for image processing and image analysis based on libraries and contextual databases within command, control, communications, computers, and intelligence (C4I) systems/ environment.