Target tracking data association method and system, storage medium and electronic equipment

Abstract

The invention provides a target tracking data association method and system, a storage medium and electronic equipment. The target tracking data association method comprises the following steps: acquiring a first feature vector of measurement data and a second feature vector of a track; screening features in the first feature vector and the second feature vector to obtain a first feature vector and a second feature vector after screening; obtaining a difference value between the screened first feature vector and the screened second feature vector as a difference value vector; and processing the difference vector by using a correlation model to judge whether the measurement data and the track are correlated or not. The target tracking data association method can utilize the association model to automatically associate the measurement data and the track.

Description

Technical Field

[0001] The present application belongs to the technical field of target tracking, and relates to a data association method, and in particular to a target tracking data association method, system, storage medium and electronic device. Background Art

[0002] Multi-target tracking refers to the process of simultaneously identifying and tracking multiple targets in a complex environment, aiming to obtain the position and motion information of these targets in real time. Multi-target tracking technology has been widely used in unmanned driving, military monitoring, traffic control, sports analysis and other fields. An important step in the multi-target tracking process is data association, which refers to matching and associating measurement data from sensors or radars with existing tracks to determine which target each measurement data should belong to. In multi-target tracking, data association can provide accurate input data for the subsequent track update, ensuring that the model can correctly track the dynamic changes of the target and effectively distinguish different targets in a complex environment. However, the existing data association methods still have some shortcomings. Summary of the invention

[0003] The embodiments of the present application provide a target tracking data association method, system, storage medium and electronic device for automatically associating measurement data and tracks using an association model.

[0004] In a first aspect, an embodiment of the present application provides a target tracking data association method, which includes: obtaining a first eigenvector of measurement data and a second eigenvector of a track; filtering features in the first eigenvector and the second eigenvector to obtain the filtered first eigenvector and the second eigenvector; obtaining the difference between the filtered first eigenvector and the second eigenvector as a difference vector; and processing the difference vector using an association model to determine whether the measurement data and the track are associated.

[0005] In an implementation of the first aspect, the training method of the association model includes: obtaining training data and its true value, the training data including feature vectors of measurement data for training and feature vectors of tracks for training; constructing positive and negative samples using the training data and its true value; and training the association model using the positive and negative samples to obtain a trained association model and effective features.

[0006] In an implementation of the first aspect, obtaining the true value of the training data includes: S1, selecting part of the training data from the training data, obtaining the true value of the part of the training data and using the part of the training data and its true value to train the machine learning model, and executing step S2; S2, selecting part of the unused training data from the training data, and executing step S3; S3, using the machine learning model after the last training to process the part of the unused training data to obtain a corresponding processing result, and executing step S4; S4, checking and correcting the processing result to obtain the true value of the part of the unused training data; if there is unused training data, jump to step S2, otherwise terminate the loop.

[0007] In an implementation of the first aspect, the complexity of the machine learning model is higher than that of the association model.

[0008] In an implementation of the first aspect, filtering features in the first feature vector and the second feature vector to obtain filtered first feature vector and second feature vector includes: filtering features in the first feature vector and the second feature vector based on the valid features to obtain the filtered first feature vector and second feature vector.

[0009] In an implementation of the first aspect, the first feature vector and the second feature vector both include a target feature and a global feature, the target feature is a feature used to describe tracking target information, and the global feature is a feature used to describe vehicle information and environment information.

[0010] In an implementation manner of the first aspect, the association model includes a support vector machine model or a decision tree model.

[0011] In a second aspect, an embodiment of the present application provides a target tracking data association system, which includes: a feature vector acquisition module, used to obtain a first feature vector of measurement data and a second feature vector of a track; a feature vector screening module, used to screen features in the first feature vector and the second feature vector to obtain the screened first feature vector and the second feature vector; a difference vector acquisition module, used to obtain the difference between the screened first feature vector and the second feature vector as a difference vector; and an association judgment module, used to process the difference vector using an association model to determine whether the measurement data and the track are associated.

[0012] In a third aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the target tracking data association method described in any one of the first aspects of the embodiments of the present application is implemented.

[0013] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising: a memory storing a computer program; and a processor communicatively connected to the memory, for executing the target tracking data association method described in any one of the first aspects of the embodiments of the present application when the computer program is called.

[0014] The data association method provided in the embodiment of the present application can automatically associate measurement data and tracks through an association model, thereby effectively overcoming the shortcomings of manually designing association rules.

[0015] In the embodiment of the present application, the association model can be used to automatically perform feature screening, thereby removing unnecessary features in the feature vector, which is beneficial to reducing the amount of calculation of the association model.

[0016] In the embodiment of the present application, a semi-automatic labeling method combining automatic labeling and manual labeling can be adopted to gradually reduce the workload of manual labeling through a closed-loop iteration, thereby improving the efficiency of true value labeling. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 Shown is a flowchart of a target tracking data association method provided in an embodiment of the present application.

[0018] Figure 2 Shown is a flowchart for training the association model in an embodiment of the present application.

[0019] Figure 3A and Figure 3B Shown is a schematic diagram of obtaining the true value of training data in an embodiment of the present application.

[0020] Figure 4 Shown is a schematic diagram of the association model training stage in an embodiment of the present application.

[0021] Figure 5 Shown is a schematic diagram of the association model deployment stage in an embodiment of the present application.

[0022] Figure 6 Shown is a schematic diagram of the structure of a target tracking data association system provided in an embodiment of the present application.

[0023] Figure 7 Shown is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0024] Component number description

[0025] 600 Target Tracking Data Association System

[0026] 601 Feature Vector Acquisition Module

[0027] 602 Feature Vector Screening Module

[0028] 603 Difference vector acquisition module

[0029] 604 Association judgment module

[0030] 700 Electronic equipment

[0031] 701 Processor

[0032] 702 Memory

[0033] 7021 Operating System

[0034] 7022 Applications

[0035] 703 Network Interface

[0036] 704 bus system

[0037] 705 User Interface

[0038] Steps S11 to S14

[0039] Steps S21 to S23

[0040] Steps S31 to S34 DETAILED DESCRIPTION

[0041] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0042] It should be noted that the illustrations provided in the following embodiments are only used to illustrate the basic concept of the present application in a schematic manner, and therefore the illustrations only show components related to the present application rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0043] In the embodiments of the present application, the words "exemplary" or "for example" represent examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0044] In the embodiments of the present application, "at least one" refers to one or more, and "plurality" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can represent: a, b, c, ab, ac, bc or abc, where a, b, c can be single or multiple.

[0045] Data association is an important step in the multi-target tracking process. The data association process includes: calculating the similarity between the measurement data and the track. Assuming there are M measurement data and N tracks, the calculated result is an M*N similarity matrix; according to the similarity matrix, the algorithm gives the measurement data that matches each track. Exemplarily, the similarity between the measurement data and the track can be obtained by calculating the distance (such as Euclidean distance) or similarity coefficient (such as cosine similarity) between the measurement data and the track.

[0046] In the process of calculating similarity, both the measurement data and the track are represented by vectors. These vectors usually contain multiple features, such as position, size, speed, texture, color, etc. In some technical solutions, these features play exactly the same role in the process of similarity calculation, that is, in these technical solutions, the importance of these features is considered to be the same. However, in practical applications, the importance of features of data from different sensors is different. For example, the distance measured by lidar is more accurate, so the features of position and size are relatively more important when calculating similarity. For another example, the camera can take color photos, so the features of texture and color are relatively more important when calculating similarity. For another example, the millimeter wave radar can directly measure the speed of the target, so the speed of the target is relatively more important when calculating similarity. However, all of the above rules need to be designed manually. In a multi-sensor fusion system, developers need to experiment with a large number of different rule combinations. The whole process is very time-consuming and there is no guarantee that the optimal rule will be found.

[0047] After obtaining the similarity matrix, the algorithm also needs to match the measurement data with the track according to the similarity. This process can be achieved through a variety of methods, such as the nearest neighbor method or Hungarian matching. However, in actual applications, not all measurement data can be matched with the track, and some measurement data may be false positives given by the perception model. At this time, the algorithm needs to define a similarity threshold to filter these false positives to ensure the correctness of the association. The threshold needs to be adjusted according to the actual application situation. For example, in an autonomous driving application, when the vehicle turns, the error in steering angle estimation will cause a large position deviation between the measurement data and the track. At this time, a relatively large similarity threshold needs to be set. In addition, for different types of targets, such as vehicles and pedestrians, the similarity thresholds are also different. If it relies entirely on manual design, there will be great difficulties, and the setting of the threshold cannot adapt well to the actual situation.

[0048] At least to address the above problems, the embodiment of the present application provides a target tracking data association method. The technical solution in the embodiment of the present application will be described in detail below in conjunction with the accompanying drawings in the embodiment of the present application.

[0049] Figure 1 The flowchart of the target tracking and association method provided by the embodiment of the present application is shown. Figure 1 As shown, the target tracking and association method provided in the embodiment of the present application includes the following steps S11 to S14.

[0050] S11, obtaining a feature vector of the measurement data as a first feature vector, and obtaining a vector of the track as a second feature vector. Both the first feature vector and the second feature vector contain multiple features.

[0051] S12, filtering the features in the first feature vector and the second feature vector to obtain filtered first feature vector and second feature vector.

[0052] Taking the first eigenvector as an example, the first eigenvector contains multiple features, however, not all features are conducive to data association. If these features are not screened and all features are directly used for calculation, it will cause a waste of computing resources. In addition, the noise present in the features will also reduce the credibility of the similarity calculation. To address this problem, in step S12, the features in the first eigenvector can be screened, and the relatively unimportant features in the first eigenvector can be deleted to obtain the screened first eigenvector. The second eigenvector is screened similarly and will not be described in detail here.

[0053] S13, obtaining the difference between the first eigenvector and the second eigenvector after screening as a difference vector.

[0054] S14, using the association model to process the difference vector to determine whether the measurement data and the track are associated. The association model is a trained machine learning model, whose input is the difference vector and output is the association judgment result of the measurement data and the track. For example, the association model output 1 indicates that the measurement data and the track are associated, and the association model output -1 indicates that the measurement data and the track are not associated.

[0055] See also Figure 2 In some implementations, the training method of the association model may include the following steps S21 to S23.

[0056] S21, obtaining training data and its true value, the training data including feature vectors of measurement data for training and feature vectors of tracks for training. Specifically, for the feature vectors of measurement data for training and feature vectors of tracks for training, the true value data is used to represent the true correlation relationship between the two types of feature vectors.

[0057] S22, construct positive and negative samples using training data and their true values.

[0058] S23, using positive and negative samples to train the association model to obtain a trained association model and effective features.

[0059] When training the association model, a large amount of high-quality true value data is required. If the true value data is obtained by manual annotation, it will take a lot of time and effort. For this problem, please refer to Figure 3A and Figure 3B The process of obtaining the true value of the training data in the embodiment of the present application includes the following steps S31 to S34.

[0060] S31, select part of the training data from the training data, obtain the true value of the part of the training data and use the part of the training data and its true value to train the machine learning model, and execute step S32. The true value of the part of the training data can be obtained by manual annotation, but is not limited to this.

[0061] S32, select some unused training data from the training data, and execute step S33.

[0062] S33, use the machine learning model trained last time to process the portion of unused training data to obtain a corresponding processing result, and execute step S34.

[0063] S34, check and correct the processing result to obtain the true value of the unused training data; if there is unused training data, jump to step S32, otherwise terminate the loop. The checking and correction of the processing result can be achieved manually, but the present application is not limited to this.

[0064] As described above, the embodiment of the present application provides a semi-automatic annotation method, which combines automatic annotation and manual annotation to improve the efficiency of obtaining true values. Specifically, small batches of true value data are manually annotated, and these data are used to train the machine learning model. The trained machine learning model can be used to process another part of the training data to determine the relevance, and the annotator only needs to check and correct these automatically annotated data (that is, the data output by the machine learning model this time) to obtain the true value data. These checked and corrected true value data can be used to train the machine learning model for the next stage. After the next stage of training is completed, the machine learning model can process the next part of the training data to determine the relevance. By continuously repeating the above iterative process, the automatic annotation ability of the machine learning model will become stronger and stronger, and the workload required for manual inspection and correction will become lower and lower. Therefore, this method can effectively improve the annotation efficiency of the true value.

[0065] In some implementations, the complexity of the above-mentioned machine learning model is higher than that of the association model. Specifically, since the above-mentioned semi-automatic annotation process is performed in an offline manner, this stage does not pay too much attention to the computational complexity and running time of the model. Therefore, a model with high complexity and high accuracy can be selected for semi-automatic annotation to improve the accuracy of the annotation, which is also conducive to reducing the number of iterations and improving the annotation efficiency. In addition, in the semi-automatic annotation process, you can also choose to use all the features in the training data, that is, do not screen the feature vectors in the training data. This method is conducive to further improving the accuracy of the annotation.

[0066] In some implementations, the machine learning model may be a support vector machine (SVM) model. Taking the linear SVM model as an example, its input is a vector, which outputs a binary value (1 or -1) after linear projection. In an embodiment of the present application, the feature vector of the measurement data used for training and the feature vector of the track used for training may be subjected to a difference operation to obtain a difference vector for training. The importance of each feature in the difference vector for training is judged by the linear projection learned by the SVM model, and then the effective features are obtained. For example, the absolute values ​​of the coefficients in the linear projection vector may be sorted, and the first L features with larger absolute values ​​may be selected as the features finally used, where L is a positive integer, and its value may be set according to actual needs.

[0067] In some implementations, the machine learning model may be a decision tree (DT) model. In an embodiment of the present application, a difference operation may be performed between the feature vector of the measurement data used for training and the feature vector of the track used for training to obtain a difference vector for training. The difference vector for training is input into the DT model, and the DT model gradually obtains the final discrimination result by performing binary classification on each feature in the vector. The construction process of the DT model is a feature selection process. Specifically, in the construction process of the DT model, more and more features will be included, and the classification error rate will become lower and lower. When the classification error rate is lower than a certain threshold, the construction process stops. At this point, the features used by the decision tree are valid features.

[0068] For example, multiple decision trees can be constructed through ensemble learning (such as Boosting or Bagging), and the union of features used by all decision trees is the effective feature.

[0069] It should be noted that the association model is a support vector machine or a decision tree, which are only two feasible methods of the embodiments of the present application, but the present application is not limited thereto. Any model that can achieve the classification task can be used as an association model, for example, a neural network can also be used as an association model in some other embodiments.

[0070] In some implementations, filtering the features in the first feature vector and the second feature vector to obtain the filtered first feature vector and the second feature vector includes: filtering the features in the first feature vector and the second feature vector based on the valid features to obtain the filtered first feature vector and the second feature vector. Taking the first feature vector as an example, the remaining features other than the valid features can be deleted therefrom to obtain the filtered first feature vector. The method for filtering the features of the second feature vector is similar and will not be repeated here.

[0071] In some implementations, the first feature vector and the second feature vector both include target features and global features. The target feature is a feature used to describe tracking target information, and the global feature is a feature used to describe vehicle information and environment information.

[0072] Exemplarily, target features may include speed, position, texture, and color of the tracked target, and global features may include motion information of the vehicle, road information, and the like.

[0073] In some implementations, the association model may be a SVM model or a DT model.

[0074] The target tracking data association method provided by the embodiment of the present application will be described in detail below through a specific example. It should be noted that the content in this example does not limit the target tracking data association method, and in specific applications, appropriate modifications can be made based on this example according to actual needs. In this example, the target tracking data association method includes an association model training phase and an association model deployment phase.

[0075] Figure 4 The diagram is shown as a schematic diagram of the training phase of the association model. Figure 4 As shown, the association model training phase includes the following steps:

[0076] S41, obtaining training data, wherein the training data includes feature vectors of measurement data used for training and feature vectors of tracks used for training, and these feature vectors include original features designed by developers.

[0077] S42, annotating the training data with true values. The true values ​​may be annotated using the semi-automatic annotation method provided in the embodiment of the present application, but the present application is not limited thereto.

[0078] S43, constructing positive and negative samples based on the training data and its true value.

[0079] S43, using positive and negative samples to train the association model to obtain a trained association model and effective features.

[0080] Figure 5 A schematic diagram showing the deployment phase of the associated model. Figure 5 As shown in the figure, the association model deployment phase includes the following steps:

[0081] S51, obtaining a first eigenvector of the measurement data and a second eigenvector of the track.

[0082] S52, filtering the features in the first feature vector and the second feature vector to obtain filtered first feature vector and second feature vector. After the feature filtering, some features originally designed manually will be deleted, and only the selected effective features will be retained.

[0083] S53, obtaining a difference vector between the filtered first eigenvector and the filtered second eigenvector.

[0084] S54, using the correlation model to process the difference vector to determine whether the measurement data and the track are correlated.

[0085] The protection scope of the target tracking data association method provided in the embodiment of the present application is not limited to the execution order of the steps listed in this embodiment. All solutions implemented by adding, reducing or replacing steps in the prior art based on the principles of the present application are included in the protection scope of the present application.

[0086] An embodiment of the present application also provides a target tracking data association system, which can implement the target tracking data association method described in the present application. However, the implementation device of the target tracking data association method described in the present application includes but is not limited to the structure of the target tracking data association system listed in the present embodiment. All structural deformations and replacements of the prior art made according to the principles of the present application are included in the protection scope of the present application.

[0087] Figure 6 is a schematic block diagram of a target tracking data association system provided by an embodiment of the present application. Figure 6 As shown, the target tracking data association system 600 includes a feature vector acquisition module 601, a feature vector screening module 602, a difference vector acquisition module 603 and an association judgment module 604. The feature vector acquisition module 601 is used to obtain the first feature vector of the measurement data and the second feature vector of the track. The feature vector screening module 602 is used to screen the features in the first feature vector and the second feature vector to obtain the first feature vector and the second feature vector after screening. The difference vector acquisition module 603 is used to obtain the difference between the first feature vector and the second feature vector after screening as a difference vector. The association judgment module 604 is used to process the difference vector using an association model to determine whether the measurement data and the track are associated.

[0088] It should be understood that the specific process of each module executing the above corresponding steps has been described in detail in the above method embodiment, and for the sake of brevity, it will not be repeated here.

[0089] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by the processor, the target tracking data association method provided by the present application embodiment is implemented. A person of ordinary skill in the art can understand that all or part of the steps in the method for implementing the above embodiment can be completed by instructing the processor through a program, and the program can be stored in a computer-readable storage medium, and the storage medium is a non-transitory medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid-state hard disk, a magnetic tape, a floppy disk, an optical disc, and any combination thereof. The above storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (DVD)), or a semiconductor medium (e.g., a solid-state disk (SSD)), etc.

[0090] An embodiment of the present application also provides an electronic device. Figure 7 is a schematic block diagram of an electronic device provided in an embodiment of the present application. Figure 7 As shown, the electronic device 700 includes: at least one processor 701, a memory 702, at least one network interface 703 and a user interface 705. The various components in the electronic device 700 are coupled together via a bus system 704. It can be understood that the bus system 704 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 704 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, the bus system 704 is not described in detail. Figure 7 In the specification, various buses are labeled as bus systems.

[0091] The user interface 705 may include a display, a keyboard, a mouse, a trackball, a click gun, keys, buttons, a touch pad or a touch screen.

[0092] It is understood that the memory 702 can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), which is used as an external cache. By way of exemplary but not limiting explanation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM). The memory described in the embodiments of the present application is intended to include but is not limited to these and any other suitable categories of memory.

[0093] The memory 702 in the embodiment of the present application is used to store various categories of data to support the operation of the electronic device 700. Examples of these data include: any executable program for operating on the electronic device 700, such as an operating system 7021 and an application 7022; the operating system 7021 includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application 7022 may include various applications, such as a media player (Media Player), a browser (Browser), etc., for implementing various application services. The target tracking data association method provided in the embodiment of the present application may be included in the application 7022.

[0094] The method disclosed in the above embodiment of the present application can be applied to the processor 701, or implemented by the processor 701. The processor 701 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of the hardware in the processor 701 or an instruction in the form of software. The above processor 701 may be a general-purpose processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The processor 701 can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor 701 may be a microprocessor or any conventional processor, etc. In combination with the steps of the accessory optimization method provided in the embodiment of the present application, it can be directly embodied as a hardware decoding processor to execute, or it can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium, which is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0095] In an exemplary embodiment, the electronic device 700 may be implemented by one or more application specific integrated circuits (ASIC), DSP, programmable logic device (PLD), complex programmable logic device (CPLD) to execute the aforementioned method.

[0096] The terms "component", "module", "system", etc. used in this specification are used to represent computer-related entities, hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program and / or a computer. By way of illustration, both applications running on a computing device and a computing device can be components. One or more components may reside in a process and / or an execution thread, and a component may be located on a computer and / or distributed between two or more computers. In addition, these components may be executed from various computer-readable media having various data structures stored thereon. Components may, for example, communicate through local and / or remote processes according to signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system and / or a network, such as the Internet interacting with other systems through signals).

[0097] Those of ordinary skill in the art will appreciate that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0098] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0099] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0100] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0101] In the above embodiments, the functions of each functional unit can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions (programs). When loading and executing computer program instructions (programs) on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center.

[0102] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk.

[0103] In summary, the present application provides a target tracking data association method, system, storage medium and electronic device. Therefore, the present application effectively overcomes the various shortcomings in the prior art and has a high industrial utilization value. The data association method provided in the embodiment of the present application can automatically associate measurement data and tracks through an association model, thereby effectively overcoming the deficiencies in the manually designed association rule method. In the embodiment of the present application, the association model can be used to automatically perform feature screening, thereby removing non-essential features in the feature vector, which is beneficial to reducing the amount of calculation of the association model. In the embodiment of the present application, a semi-automatic annotation method combining automatic annotation and manual annotation can be adopted, and the workload of manual annotation is gradually reduced by a closed-loop iteration method, thereby improving the efficiency of true value annotation. Therefore, the present application effectively overcomes the various defects in the prior art and has a high industrial value.

[0104] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A target tracking data association method, characterized in that: include: Obtaining a first eigenvector of the measurement data and a second eigenvector of the track; Screening the features in the first feature vector and the second feature vector to obtain screened first feature vector and second feature vector; Obtaining a difference between the first eigenvector and the second eigenvector after the screening as a difference vector; The difference vector is processed using a correlation model to determine whether the measurement data and the track are correlated.

2. The target tracking data association method according to claim 1, characterized in that: The training method of the association model includes: Acquire training data and true values ​​thereof, wherein the training data includes a feature vector of measurement data used for training and a feature vector of a track used for training; Constructing positive and negative samples using the training data and its true value; The association model is trained using the positive and negative samples to obtain a trained association model and effective features.

3. The target tracking data association method according to claim 2, characterized in that: Obtaining the true value of the training data includes: S1, selecting part of the training data from the training data, obtaining the true value of the part of the training data and using the part of the training data and its true value to train the machine learning model, and executing step S2; S2, selecting some unused training data from the training data, and executing step S3; S3, using the machine learning model trained last time to process the unused part of the training data to obtain a corresponding processing result, and executing step S4; S4, checking and correcting the processing result to obtain the true value of the unused training data; if there is unused training data, jump to step S2, otherwise terminate the loop.

4. The target tracking data association method according to claim 3, characterized in that: The complexity of the machine learning model is higher than that of the association model.

5. The target tracking data association method according to claim 2, characterized in that: Screening the features in the first feature vector and the second feature vector to obtain the screened first feature vector and the second feature vector includes: Features in the first feature vector and the second feature vector are screened based on the effective features to obtain the screened first feature vector and the second feature vector.

6. The target tracking data association method according to claim 1, characterized in that: The first feature vector and the second feature vector both include a target feature and a global feature, wherein the target feature is a feature used to describe tracking target information, and the global feature is a feature used to describe vehicle information and environment information.

7. The target tracking data association method according to claim 1, characterized in that: The association model includes a support vector machine model or a decision tree model.

8. A target tracking data association system, characterized in that: The target tracking data association system comprises: A feature vector acquisition module, used to acquire a first feature vector of the measurement data and a second feature vector of the track; A feature vector screening module, used for screening features in the first feature vector and the second feature vector to obtain the screened first feature vector and the second feature vector; A difference vector acquisition module, used for acquiring the difference between the first eigenvector and the second eigenvector after the screening as a difference vector; The association judgment module is used to process the difference vector using an association model to judge whether the measurement data and the track are associated.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the target tracking data association method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: The electronic device comprises: A memory storing a computer program; A processor is communicatively connected to the memory, and executes the target tracking data association method according to any one of claims 1 to 7 when calling the computer program.

Images