Method and system for quickly identifying and tracking low-slow small target

Abstract

The invention relates to a method and a system for quickly identifying and tracking a low-slow small target. The method comprises the following steps: constructing a video device array for a target area; each video device has a first state and a second state, and independently performs video stream acquisition and local processing in parallel; performing multi-view data association and three-dimensional positioning on each moving target based on the current moving target information and the historical moving target information, and determining an associated video device set of each moving target in real time; and determining the dominant video device group of each moving target in real time, and performing fine identification and tracking on the moving target based on the dominant video device group. According to the invention, early warning, real-time identification, continuous tracking and threat assessment of low, slow and small targets such as unmanned aerial vehicles, model airplanes, birds and the like are realized. The system can be widely applied to key scenes such as important place air defense, border patrol, major activity security and airport clearance protection.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent identification, and particularly relates to a method and system for quickly identifying and tracking low, slow and small targets. BACKGROUND

[0002] Low, slow and small flying targets refer to aircrafts with slow flight speed, small size, low flight altitude and difficult to be detected by military and civilian radars. The identification and tracking of low, slow and small flying targets are facing increasingly urgent development needs. The urgency is derived from the serious security challenges brought by such targets, which are easy to obtain and manipulate and have the possibility of dangerous occurrence. With the rapid development of low-altitude economy, unmanned aerial vehicle technology is widely used in many fields, such as logistics, agricultural monitoring, security and the like. A typical scenario is low, slow and small target unmanned aerial vehicles, and with their widespread use, they have brought great challenges to traditional unmanned aerial vehicle detection and countermeasure systems. These low, slow and small unmanned aerial vehicles are usually small in size, slow in flight speed, complex in flight trajectory, and often fly in complex environments such as cities and airports, which makes it difficult for existing detection technologies to effectively identify their existence, especially in the presence of a large amount of interference and occlusion, the accuracy and real-time performance of target detection often cannot meet the needs. Especially for swarms, high-resolution phased array radars and multi-transmit multi-receive radars are mainly developed. The former tries to resolve multiple tracks by rapidly scanning the airspace with agile beams, and the latter improves the resolution capability for dense targets by using its virtual aperture characteristics. However, radars still have inherent limitations: it is difficult to distinguish targets that are very close to each other, and it is impossible to identify the identity of the unmanned aerial vehicles and distinguish them from bird flocks.

[0003] However, the traditional radar detection method has difficulty in discovering targets with small radar cross section, and it is difficult to make fine classification; for the swarm, its communication mode may use frequency hopping, networking and other anti-jamming techniques, and once the swarm performs autonomous pre-programmed tasks, this technology is invalid. Single photoelectric identification method is limited by field of view and environment, and it is difficult to realize large-scale continuous monitoring. The existing identification method mainly depends on single sensor or simple multi-sensor splicing, which has disadvantages. First, the data fusion level of these methods is shallow, and the correlation problem of observation data of different sensors cannot be fundamentally solved in geometric space, resulting in high false alarm rate and frequent switching of target identity identification. Second, the allocation of computing resources is unreasonable, and the extensive mode of full-load operation of all sensors is often used, which deeply processes a large amount of invalid or low-value information, causing huge waste of computing power, and it is difficult to support large-scale and long-time defense application. More importantly, the existing system lacks the ability to perceive its own monitoring state, and cannot judge in real time whether the tracking quality is deteriorating or the association is wrong, and its monitoring process is a black box that cannot introspect, and its reliability is questionable. More importantly, in recent years, the significant reduction in the cost of image sensors, processors and communication hardware has laid a solid material foundation for the paradigm innovation of low, slow and small target identification and tracking technology; this trend directly gives rise to two key possibilities: first, the popularization of hardware cost makes large-scale and high-density camera array layout from expensive concept verification to large-scale engineering application, providing a physical premise for realizing wide-area, multi-angle and dead-angle-free collaborative perception; second, the decline in the price of computing hardware makes it possible to move computing power from central servers to the edge side of each camera node, which fundamentally overturns the traditional acquisition-transmission-centralized processing architecture; based on the above hardware condition changes, the present invention improves the traditional technology, and its core is distributed array setting and hierarchical computing front-end. The traditional scheme is limited by cost, and usually uses a small number of high-priced sensors and transmits raw data back to the central server for centralized processing. This post-computing mode inevitably introduces network delay and bandwidth bottleneck, resulting in slow system response, which is difficult to meet the needs of rapid identification and real-time tracking of low, slow and small targets. When facing a large number of small, cooperative swarms, the traditional single sensor system is easily trapped in the dilemma of missed detection and misjudgment due to narrow field of view, saturated computing power or feature confusion.

[0004] The present application makes full use of the hardware cost advantage, deploys large-scale camera array, and completely avoids the above problems by layering and grading the computing task in advance: each video device node independently and in parallel completes the local motion target detection and preliminary identification and other lightweight computing, thereby realizing partial computing in advance, uploading only effective metadata rather than massive video stream, and keeping complex computing function for the advantage video device part; then, higher-level collaborative scheduling computing is performed in the monitoring server, which greatly reduces the network transmission load, shortens the closed-loop response time from data acquisition to decision output, enables the entire system to make near-real-time response to high-speed maneuvering low, slow and small targets with unprecedented efficiency, and thereby realizes the dual optimization of performance and cost; especially in the case of low, slow and small target swarm, it shows its unique technical advantages and excellent detection ability; through the collaborative layout of large-scale distributed video device array, it realizes wide-area seamless coverage of airspace in the detection layer, and the multi-view and overlapping observation characteristics make it difficult for individual targets in the swarm to hide in blind areas, and through the almost simultaneous detection information of the array structure, the overall appearance of the swarm can be quickly perceived and confirmed, and its scale and approximate spatial distribution can be preliminarily judged. SUMMARY

[0005] In order to solve the above problems in the prior art, the present application provides a low, slow and small target rapid identification and tracking method and system, which comprises: Step S1: constructing a video device array for a target area; the video device array comprises a plurality of video devices; the field of view range of each video device in the video device array is overlapping or non-overlapping; Step S2: each video device has a first state and a second state, and independently and in parallel performs video stream acquisition and local processing; each video device is in the first state to perform real-time detection of motion targets in the field of view range; when it is one of the advantage video device group as a motion target, it enters and maintains the second state to perform high-precision identification and tracking on the motion target; Step S3: all video devices send current motion target information to the monitoring server in real time; based on the current motion target information and the historical motion target information, multi-view data correlation is performed on each motion target to determine the associated video device set of each motion target in real time; the associated video device set is a set of video devices that can effectively observe the motion target; Step S4: Real-time determination of the dominant video device group for each moving target based on the associated video device set, and sending a state switching message to the current dominant video device group and the new dominant video device group when a change in the dominant video device group is found, the corresponding video devices in the current dominant video device group closing the second state and keeping the first state for the changed moving target, and the corresponding video devices in the new dominant video device group opening the second state and keeping the first state for the changed moving target.

[0006] Further, the moving target is identified and tracked with high precision using the dominant video device group, and in the first state, each video device detects the moving target in its detection range; specifically, for a known global moving target, video devices are selected from all associated video device sets to form a dominant video device group; the dominant video device group is used for high-precision identification and tracking of the moving target; the dominant video device group includes a master dominant video device and a slave dominant video device; wherein: the master dominant video device is used for high-confidence identification and image evidence; the slave dominant video device is used to cooperate with the master dominant video device to perform stereo vision triangulation and output real-time high-precision three-dimensional coordinates.

[0007] Further, a uniform second state is set for all moving targets in each video device, and the entering and maintaining are managed; in the case of uniform setting, when the video device does not enter the second state due to any moving target, it enters the second state upon receiving the state switching message; when it first enters the second state and maintains the second state; when maintaining the second state, if the state switching message is received and the second state needs to be closed for a moving target, only the second state setting maintained by the video device for that moving target needs to be closed, while the second state setting maintained for other moving targets is preserved; if the video device maintains the second state only for one moving target, the second state can be directly closed.

[0008] Further, the second state is independently set for each moving target in each video device, and the entering and maintaining are managed; in the case of independent setting, the second state entering and maintaining associated with the moving target are performed based on the state switching message, or closed; in the case of second state maintaining, the dominant video device group performs high-precision identification and tracking of the moving target.

[0009] Further, the real-time determination of the dominant video device group for each moving target is as follows: the advantage score of each video device in the associated video device set is calculated based on the moving target size, centering degree, image clarity, and video device resolution, and the two video devices with the highest advantage scores are selected to form the dominant video device group, wherein: the one with the highest advantage score is the master dominant video device, and the one with the second highest advantage score is the slave dominant video device.

[0010] Further, wherein: the motion target information includes one or more of image coordinates, features, timestamps; the detection server performs timestamp alignment on all received motion target information.

[0011] Further, the monitoring server determines a sequence of associated video devices for each motion target based on the set of associated video devices for each motion target; and performs real-time anomaly monitoring based on changes in the sequence of associated video devices.

[0012] A low, slow and small target fast identification and tracking platform, the platform is used to realize the low, slow and small target fast identification and tracking method.

[0013] A low, slow and small target fast identification and tracking server, comprising a processor, the processor and a memory are coupled, the memory stores program instructions, when the program instructions stored in the memory are executed by the processor, the low, slow and small target fast identification and tracking method is realized.

[0014] A low, slow and small target fast identification and tracking system, the system is used to realize the low, slow and small target fast identification and tracking method.

[0015] A computer readable storage medium, comprising a program, when it runs on a computer, makes the computer execute the low, slow and small target fast identification and tracking method.

[0016] The beneficial effects of the present application include: (1) By constructing a redundant full coverage range through a video device array, partial calculation is pre-posed, only effective metadata but not massive video streams are uploaded, and complex calculation functions are maintained for advantageous video devices, each video device works in a first state and a triggerable second state, fine classification is differentiated for each motion target, and each motion target is independently maintained with an associated video device set and an advantageous video device group, thereby providing differentiated customized identification and tracking in the overall low, slow and small target fast identification and tracking process without significantly increasing hardware and software overhead, and providing a basis for differentiated parameter setting; (2) Fully utilizing big data information, multi-view data association and three-dimensional positioning are performed on each motion target based on current motion target information and historical motion target information, the associated video device set of each motion target is determined in real time, and based on the associated video device set, anomaly monitoring is supported based on simple quantitative calculation, motion targets that do not meet expected trajectory changes, disappearances and appearances within the target range are sensitively discovered, and the robustness of identification is supplemented. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings: Figure 1 A schematic diagram of the low, slow and small target fast identification and tracking method provided by the present application. DETAILED DESCRIPTION

[0018] The present application will be described in detail below with reference to the drawings and specific embodiments, wherein the illustrative embodiments and descriptions are only used to explain the present application and do not constitute limitations on the present application.

[0019] As shown in the accompanying Figure 1 The method comprises the following steps: Step S1: constructing a video device array for a target area; the video device array comprises a plurality of video devices; the field of view of each video device in the video device array is overlapped or not overlapped; the field of view of all video devices in the video device array covers the target area; Preferably, the video device array and each video device therein are calibrated and initialized; specifically, the intrinsic parameters of each video device are calibrated to determine the focal length, principal point coordinates, distortion coefficient and other intrinsic parameter matrix; the extrinsic parameters of the video device array are calibrated; Preferably, the intrinsic parameter calibration is completed using a calibration board such as a checkerboard; the extrinsic parameter calibration of the video device array is performed by placing one or more obvious calibration objects or moving one or more feature points in the common field of view of all video devices, that is, in the target area, calculating the multi-view correspondence of each relative to the calibration object, deriving the view angle relationship between the video devices, selecting a fixed geographic coordinate point as the origin of the world coordinate system, converting the extrinsic parameters of all video devices to the coordinate system, and further using a rotation matrix and a translation vector to represent the relative position and attitude relationship between the video devices to complete the extrinsic parameter calibration; Step S2: each video device has a first state and a second state, and independently and in parallel performs video stream acquisition and local processing; each video device is in the first state to perform real-time detection of moving targets in the field of view; when it is in one of the advantage video device groups of a specific moving target, it simultaneously enters and remains in the second state to perform high-precision identification and tracking of the specific moving target; the advantage video device group of the specific moving target comprises the video device that enters and remains in the second state; Preferably, each video device uses the intrinsic parameters obtained by calibration to perform real-time lens distortion correction to ensure image geometric accuracy; uses a background segmentation model to extract foreground moving pixel blocks; clusters and filters the foreground blocks to obtain candidate regions of moving targets; and inputs the candidate regions into a coarse classification module to obtain coarse classification of moving targets; Preferably: the coarse classification module only needs to judge whether the moving target is a low, slow and small target; the low, slow and small target determined by classification contains predetermined types such as birds, aircrafts and unmanned aerial vehicles, and does not need to be finely classified; for the confirmed moving target, its two-dimensional image coordinates, appearance features, time stamp and video device ID are extracted as the moving target information; Step S3: multi-view data association and three-dimensional positioning are performed on each moving target based on the current moving target information and the historical moving target information, and a set of associated video devices of each moving target is determined in real time; the set of associated video devices is a video device capable of effectively observing the moving target; specifically: all video devices send the moving target information to the monitoring server in real time; the monitoring server performs time synchronization, alignment and data association on the moving target information; wherein: the moving target information includes image coordinates, features, time stamps and the like; the time stamp alignment performed on all moving target information can be performed for network delay and video array node calibration; The multi-view data association specifically includes: geometric constraint association and / or appearance feature association; a unique identifier is set for each moving target after association; The geometric constraint association specifically includes: using the calibrated external parameters, the image coordinates of the target in the video device A are reversely projected into the coordinate system to form an observation ray, and it is judged whether the observation ray passes through the target point near the imaging plane of the video device B; if the observation rays of multiple video devices for the same moving target intersect at a point (a local range) in the spatial coordinate system, the association is successful, and the multiple video devices are set as the set of associated video devices of the same moving target; the set of associated video devices is a video device capable of effectively observing the moving target; using the moving target information of the set of associated video devices and stereo vision triangulation, the three-dimensional coordinates (X, Y, Z) of the moving target in the coordinate system are solved; the length of the observation ray or the two-dimensional image coordinate information of the moving target can be used to screen the set of associated video devices, so that the observation ray with smaller length or the moving target with larger two-dimensional image coordinate span (motion target frame size) enters the set of associated video devices of the moving target; Further: time-based geometric constraint screening is performed, specifically: the historical three-dimensional coordinates of the moving target are used to predict the position coordinates, the next time position coordinates are obtained as the predicted position, and the predicted position is projected onto the image plane of each video device in the set of associated video devices to obtain the predicted pixel coordinates; for each associated video device, the predicted pixel coordinates and the current two-dimensional image coordinates are compared, and when the distance between the two is less than a distance threshold, the associated video device is kept in the set of associated video devices of the moving target, otherwise it is deleted from the set of associated video devices; the distance threshold can be set relatively leniently here; The appearance feature correlation is specifically: calculating cosine similarity or Euclidean distance between appearance feature vectors of moving targets detected by different video devices, and when similarity obtained by multiple video devices is smaller or distance is smaller, the correlation is successful, and the multiple video devices are set as a set of correlated video devices of the same moving target; the appearance feature vector of the moving target can be obtained by the video device in the first state or further calculated based on two-dimensional image coordinates of the moving target by the monitoring server; Alternatively, when the correlation is based on geometric constraints and appearance features at the same time, JPDA joint probability data association is performed to improve the correlation accuracy when occlusion or field of view edge occurs; Step S4: Real-time determination of the dominant video device group of each moving target, when the dominant video device group changes, a state switching message is sent to the current dominant video device group and the new dominant video device group, the corresponding video devices in the current dominant video device group turn off the second state and keep the first state, and the corresponding video devices in the new dominant video device group turn on the second state and keep the first state; Obviously, the corresponding video devices are part of the change; That is, the dominant video device group is used for high-precision identification and tracking of the moving target, and each video device detects the moving target in its detection range; Specifically: for a known global moving target, video devices are selected from all correlated video device groups to form a dominant video device group; the dominant video device group is used for high-precision identification and tracking of the moving target; the dominant video device group includes master dominant video devices and slave dominant video devices; wherein: the master dominant video device is used for high-confidence identification and image evidence; the slave dominant video device is used for stereovision triangulation with the master dominant video device, and real-time output of high-precision three-dimensional coordinates (X, Y, Z); Obviously, the second state can be independently set for each moving target and managed for its entry and retention, or a unified second state can be set for all moving targets and managed for entry and retention; in the case of independent setting, the second state entry and retention associated with the moving target are based on the state switching message; in the case of unified setting, when the video device does not enter the second state due to any moving target, the second state is entered when the state switching message is received, and the second state is entered for the first time and kept; when keeping the second state, if the state switching message is received and the second state needs to be closed, only the second state setting kept by the video device due to the moving target needs to be closed, while the second state setting kept due to other moving targets is preserved, if the video device keeps the second state only due to one moving target, the second state can be directly closed; each moving target can be set and managed by a state list corresponding to the moving target; Alternatively, the master and slave advantage video devices are used for high-confidence recognition and image evidence, and high-precision three-dimensional coordinates; the master and slave advantage video devices are redundant devices; Alternatively, the advantage video device group contains three advantage video devices, all of which are used for high-confidence recognition and image evidence, and high-precision three-dimensional coordinates; the recognition results of the three are voted to determine the final recognition result; The advantage video device group of each moving target is determined in real time, specifically: the advantage score of each video device in the associated video device group is calculated based on the size, centering degree, image clarity and video device resolution of the moving target, and the two video devices with the highest advantage scores are selected to form the advantage video device group, wherein the video device with the highest advantage score is the master advantage video device, and the video device with the second highest advantage score is the slave advantage video device; the video frames of all video devices are monitored in real time, and the advantage video device group is re-determined when the determination condition is met, and when the advantage video device group combination changes, a state switching message is sent to the current advantage video device group combination and the new advantage video device group combination, the corresponding video devices in the current advantage video device group combination turn off the second state and keep the first state, and the corresponding video devices in the new advantage video device group combination turn on the second state and keep the first state; the determination condition is that the preset frame number is reached or the moving target approaches the edge of the field of view of the corresponding video device in the advantage video device group combination; it can be seen that the first state is the resident state of the video device; The advantage score of each video device k in the associated video device group is calculated based on the size, centering degree, image clarity and video device resolution of the moving target , specifically: the size score , the centering score and the clarity score are calculated based on the following formulas (1)-(3) , and the resolution score is determined; linear or nonlinear normalization is performed to obtain the corresponding normalized value , , ; the advantage score is calculated based on the normalized value and formula (4); wherein: is the score coefficient, which is a preset value, the sum of the values is 1, and can be adjusted according to the specific application scenario to improve the recognition accuracy and , and to improve stability , is a relatively fixed static value; and are the frame width and frame height of the moving target in the video device k, respectively; is a size threshold of the moving target, which is a preset value; is a parameter for controlling the decay speed of the score, which is an adjustable hyperparameter; and are respectively the center point coordinates of the bounding box range of the moving target; and are respectively the center point coordinates of the bounding box range of the image; , , are respectively the sharpness, brightness and contrast, is the resolution; (1); (2); (3); (4); The advantage video device group is used for high-precision identification and tracking of the moving target; specifically, the main advantage video device uses an accurate identification model to finely classify the moving target and track based on the fine classification result; a high-definition video stream of the tracking process is recorded, and the clearest target close-up image is intercepted for subsequent evidence collection and analysis; for example, using a deep learning model such as YOLOv8, Faster R-CNN for fine classification to identify that it is a DJI Mavic 3, not just a drone; Preferably, the method further comprises: the monitoring server performs abnormal monitoring based on the change of the associated video device set cls of each moving target; specifically, the change degree of the set is calculated based on the following formula (5) When the change degree of the set is greater than the change degree threshold, it is determined that an abnormality has occurred; after the abnormality occurs, an alarm is given; the monitoring server responds to the abnormality alarm and switches all video devices in the associated video device set of the moving target to enter the second state; wherein: cls is the current associated video device set, becls is the previous associated video device set; is the number of elements of the set, that is, the size of the set; the change degree threshold is a preset value; for example, set to 5-50%; of course, the setting of this threshold is related to the target area and the size of the video device array, the relationship between each video device and the size of the video device array; it is also related to the sensitivity of the abnormal alarm; (5); In order to further improve the sensitivity of abnormal monitoring, the method further comprises: the monitoring server determines the associated video device sequence of each moving target based on the associated video device set of each moving target; and performs real-time abnormal monitoring based on the change of the associated video device sequence; The association video device set based on each motion target determines an association video device sequence of each motion target, specifically: arranging each video device in the association video device set from large to small based on the size of the advantage score to form an association video device sequence; The real-time change of the association video device sequence is monitored for abnormalities; specifically: vectorizing the association video device sequence to construct a feature vector, each element in the feature vector corresponding to a fixed video device in the video array; placing the position of the video device in the association video device sequence into the corresponding position in the feature vector to obtain the instantiated feature vector; the missing value is set to a default value, such as the maximum value or 0; the missing value indicates that the video device does not appear in the current association video device sequence; calculating the sequence order change degree and the sequence entropy based on the feature vector, and monitoring for abnormalities based on the sequence order change degree and the discrete degree; The sequence order change degree and the discrete degree are monitored for abnormalities; specifically: calculating a first sequence order change degree reflecting the overall structural change degree of the sequence and a second sequence order change degree reflecting the change discrete degree, and monitoring for abnormalities based on the first sequence order change degree and / or the second sequence order change degree; The first sequence order change degree is calculated , specifically: calculating the first sequence order change degree based on the following formulas (6)-(8) ; wherein: is the current feature vector, is the previous feature vector; is the monitoring period; k is the video device number, ; When ρ = 1, it indicates no change; when ρ = -1, it indicates complete reversal; when ρ ≈ 0, it indicates no monotonic relationship; when ρ = 1, = 0, when ρ = -1, =1; obviously, the closer to 0, the smaller the order change degree; the closer to 1, the larger the order change degree; K is the number of video devices; it can also be the maximum value of the identifier; (6); (7); (8); The second sequence order change degree is calculated , specifically: calculating the second sequence order change degree based on the following formulas (9)-(10) ; (9); ​ (10); monitoring server switches all video devices in the associated video device set of the moving target into the second state in response to the anomaly; the change degree threshold value can be determined based on historical data and preset; for example, the first change degree threshold value is set to 0.05-0.3; the second change degree threshold value is set in relation to the arrangement of the video device array, the field of view size, the moving speed of the small target, etc., and can be modulated to a unified, experience-based relative threshold value (such as 0.4-0.7) based on simulation data for judgment; Preferably, the change degree threshold value is determined based on the identification result of the moving target by the dominant video device group; when the identification result of the moving target shows that the moving speed and the change degree of the moving track are large, a relatively large change degree threshold value is set; otherwise, a smaller change degree threshold value can be set; that is, specific parameters in the identification device and the identification strategy can be customized for each moving target, thereby providing differentiated customized identification and tracking in the overall low, small and slow target fast identification and tracking process without significantly increasing the hardware and software overhead; in addition, when it is found that the moving target does not meet the expected track change, disappearance and appearance within the target range, there will be a dramatic change in the sequence or set, thereby supporting anomaly monitoring based on simple quantitative calculation based on the associated video device set, and sensitive discovery of the situation that the moving target does not meet the expected track change, disappearance and appearance within the target range, thereby supplementing the robustness of identification; of course, a stepped threshold value judgment can be provided for the change degree and the change degree to grade different anomaly levels, which will not be described here; Preferably, the dominant video device group of each moving target is determined in real time based on a first period; anomaly monitoring is performed based on the change of the associated video device sequence in real time based on a second period; wherein the first period is less than or equal to the second period; More importantly, for the swarm of low, small and slow targets, the dynamic dominant video device set and the combination strategy adopted by the system make it naturally suitable for the high dynamic and multi-target computing environment of the swarm; the processing capability of each video device node can process multiple targets in the field of view in parallel, greatly relieving the computing pressure of the central monitoring server; the optimal computing resources can be intelligently allocated to different subsets or key individuals in the swarm, ensuring that the most sufficient perception and the most accurate tracking can be obtained in the area with the highest swarm density, thereby realizing multi-granularity and integrated monitoring of the overall macro behavior of the swarm and the micro motion of the individuals, and providing solid and comprehensive information support for subsequent early warning and decision response; Based on the same inventive concept, the application further provides a low, slow and small target rapid identification and tracking system, which is used to realize the low, slow and small target rapid identification and tracking method; the system comprises a monitoring server and a video device array; the video device array comprises a plurality of video devices, and each video device is in communication connection with the monitoring server through wired or wireless mode; Based on the same inventive concept, the application further provides a low, slow and small target rapid identification and tracking server, which is provided with the monitoring server and is used to realize the low, slow and small target rapid identification and tracking method; the monitoring server is a cloud server; Based on the same inventive concept, the application further provides a low, slow and small target rapid identification and tracking device, which is used to realize the low, slow and small target rapid identification and tracking method; Based on the same inventive concept, the application further provides a low, slow and small target rapid identification and tracking platform, which is used to realize the low, slow and small target rapid identification and tracking method; The system realizes accurate association and fusion of observation data in three-dimensional space through accurate space-time calibration and multi-view geometric constraint, fundamentally ensures the uniqueness of target identity and the continuity of trajectory, greatly reduces the probability of false association, introduces a dynamic advantage camera selection strategy, only a few cameras with the best view undertake the task of fine identification and positioning with high load, and other cameras are in a low-power guard state, thereby realizing the multiplication of calculation efficiency, making it possible to deploy a large-scale camera array, finally, the system has strong self-perception and self-diagnosis capability, converts abstract associated video device sequences into quantifiable feature vectors, and performs real-time time series analysis, the system can sensitively capture abnormal changes of the tracking link, realize the leap from passive monitoring to active early warning, and significantly improve the reliability and robustness of the system; in summary, the method realizes a high-efficiency, reliable and scalable low, slow and small target identification and tracking solution through deep fusion, intelligent collaboration and resource optimization; A computer program, which can also be referred to or be included in a computer program product, can be written in any form of programming language, including compiled or interpreted languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program can, but need not, correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub programs, or portions of code). A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and are interconnected by a communication network.

[0020] Those skilled in the art will appreciate that embodiments of the present application can be devised for a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0021] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing unit or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 The flowchart illustrations and / or block diagrams Figure 1 Means for carrying out any one or more of the functionality

[0022] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 The flowchart illustrations and / or block diagrams Figure 1 Means for carrying out any one or more of the functionality

[0023] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0024] Finally, it should be noted that the above-mentioned embodiments are merely used to illustrate the technical solutions of the present application, rather than limit the technical solutions of the present application. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered in the protection scope of the claims of the present application.

Claims

1. A method for fast identification and tracking of small, slow targets, characterized in that, The method includes: Step S1: Construct a video device array for the target area; the video device array contains multiple video devices; the field of view of each video device in the video device array may or may not overlap; Step S2: Each video device has a first state and a second state, and performs video stream acquisition and local processing independently and in parallel; each video device is stationary in the first state and performs real-time detection of moving targets within its field of view; when it is one of the dominant video device groups for a moving target, it enters and maintains the second state to perform high-precision identification and tracking of the moving target. Step S3: All video devices send the current moving target information to the monitoring server in real time; based on the current moving target information and historical moving target information, multi-view data association is performed on each moving target to determine the associated video device set for each moving target in real time; the associated video device set is the set of video devices that can effectively observe the moving target. Step S4: Based on the associated video device set, determine the dominant video device group for each moving target in real time. When a change in the dominant video device group is detected, send a state switching message to the current dominant video device group and the new dominant video device group. The corresponding video device in the current dominant video device group closes the second state and maintains the first state for the changed moving target. The corresponding video device in the new dominant video device group opens the second state and maintains the first state for the changed moving target.

2. The method for fast identification and tracking of low-speed, small targets according to claim 1, characterized in that, The moving target is identified and tracked with high precision using a group of dominant video devices. In the first state, each video device detects moving targets within its detection range. Specifically, for a known global moving target, video devices are selected from all associated video devices to form a group of dominant video devices. This group of dominant video devices is used for high-precision identification and tracking of the moving target. The group of dominant video devices includes a primary dominant video device and secondary dominant video devices. The primary dominant video device is used for high-confidence identification and image forensics. The secondary dominant video devices are used in conjunction with the primary dominant video device to perform stereoscopic vision triangulation and output high-precision three-dimensional coordinates in real time.

3. The method for fast identification and tracking of low-speed, small targets according to claim 2, characterized in that, In each video device, a unified second state is set for all moving targets, and entry and maintenance are managed. Under the unified setting, when the video device enters the second state upon receiving a state switching message, it first enters the second state and maintains it. When maintaining the second state, if a state switching message is received and it is necessary to close the second state for a moving target, only the second state setting maintained by the video device for that moving target needs to be closed, while the second state settings maintained by other moving targets are retained. If the video device maintains the second state only for one moving target, the second state can be closed directly.

4. The method for fast identification and tracking of small, slow targets according to claim 2, characterized in that, Each video device independently sets a second state for each moving target and manages its entry and maintenance. When set independently, the second state associated with the moving target is entered or maintained, or closed, based on the state switching message. When the second state is maintained, the superior video device group performs high-precision identification and tracking of the moving target.

5. The method for fast identification and tracking of slow, small targets according to claim 4, characterized in that, The method for determining the dominant video device group for each moving target in real time is as follows: based on the size of the moving target, the centering degree, the image clarity and the resolution of the video device, the advantage score of each video device in the associated video device is calculated, and the two video devices with the highest advantage scores are selected to form the dominant video device group, wherein: the one with the highest advantage score is the primary dominant video device, and the one with the second highest advantage score is the secondary dominant video device.

6. The method for fast identification and tracking of small, slow targets according to claim 5, characterized in that, in: The moving target information includes one or more of image coordinates, features, and timestamps; the detection server performs timestamp alignment on all received moving target information.

7. The method for fast identification and tracking of small, slow targets according to claim 5, characterized in that, The monitoring server determines the sequence of associated video devices for each moving target based on the set of associated video devices for each moving target; and performs real-time anomaly monitoring based on changes in the sequence of associated video devices.

8. A server for fast identification and tracking of small, slow targets, characterized in that, The method includes a processor coupled to a memory, the memory storing program instructions, which, when executed by the processor, implement the fast identification and tracking method for slow, small targets as described in any one of claims 1-7.

9. A fast identification and tracking system for low-speed, small targets, characterized in that, The system is used to implement the fast identification and tracking method for slow, small targets as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, Includes a program that, when run on a computer, causes the computer to perform the method for fast identification and tracking of slow, small targets as described in any one of claims 1-7.

Images