Nonlinear scene multi-target tracking method and system and storage medium

Abstract

The invention discloses a non-linear scene multi-target tracking method and system and a storage medium, and the method comprises the steps: constructing a non-linear motion model and a measurement model, and specifically, representing a non-linear state process and a measurement process through a non-linear model; according to multi-target tracking in a two-dimensional plane, determining a motion equation of each target, setting the motion equation as a collaborative turning model, and determining a motion equation of the target; constructing a system state transition matrix, a process noise distribution matrix and a measurement equation; according to the nonlinear motion model and the measurement model, extraction of multi-target state estimation is carried out through an FAUKF-PHD filter, a state vector of each target is obtained, and tracking of each target is completed; and verifying the tracking result of each target based on the multiple pieces of preset scene information. The method is high in tracking precision and low in cost, can improve the effectiveness of the method, and can be widely applied to the technical field of computers.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a nonlinear scene multi-target tracking method, system and storage medium. Background Art

[0002] Holographic staring radar is a multifunctional radar that covers the entire airspace at the same time. Its characteristics are that it uses a wide beam for transmission and uses digital beam forming to generate multiple narrow beams to cover the transmission area. The application of low-gain wide beam will also weaken the received target echo signal strength accordingly, so it is necessary to adopt a long-term accumulation method to improve the detection probability of the target. The long-term accumulation function enables the holographic staring radar to obtain higher gain and Doppler resolution, and can effectively separate moving targets from clutter. Like traditional radars, target tracking is also a classic problem in holographic radar monitoring. How to track multiple targets in real time and accurately in a chaotic tracking environment while shielding the interference of external factors such as clutter is a difficult and key issue in the current target tracking field. The two mainstream multiple target tracking (MTT) technologies are the multiple target tracking method based on data association (DA) and the multiple target tracking method based on random finite set (RFS). The MTT method based on data association transforms the multi-target tracking problem into multiple single-target tracking problems, including classic algorithms such as Multiple Hypothesis Tracking (MHT) and Joint Probabilistic Data Association (JPDA). Such algorithms can achieve good tracking results in scenarios with little environmental clutter, a certain number of targets, and a high signal-to-noise ratio. However, when dealing with multi-target problems with relatively complex scenarios, data association is prone to errors, resulting in a decrease in tracking performance. Moreover, when the number of targets and the number of false alarms are large, the amount of computation increases exponentially. The MTT method based on random finite sets proposed by Mahler transforms the multi-target tracking problem into a set-valued estimation problem, and uses the probability hypothesis density function (PHD) to describe the state value and observation value of the target, so that multi-target state estimation under the Bayesian framework can be achieved. Compared with data association algorithms, it can avoid the explicit association between state and measurement, save the complex data association calculation, and significantly improve the effectiveness. As an emerging multi-target tracking method, it has gradually become a research hotspot in the field of MTT.

[0003] PHD filters usually cannot obtain closed-form solutions because of the difficult-to-calculate integrals involved in their iterative process. In order to obtain a closed-form solution, Vo et al. proposed two main approximation methods, namely Sequential Monte Carlo PHD (SMC-PHD) filter and Gaussian Mixture PHD (GM-PHD) filter. The nonlinear fitting characteristics of the former make it suitable for nonlinear non-Gaussian environments, but the SMC-PHD filter has the disadvantages of particle degradation and difficulty in extracting multi-target states. In view of the high real-time performance requirements in the radar multi-target tracking process, when the filter increases the number of particles to improve the tracking accuracy of multiple targets, it will seriously reduce the computational efficiency and is not conducive to engineering implementation. The latter represents the target's existence probability distribution as a Gaussian mixture model. The update and prediction of the probability density function can be completed by calculating the Gaussian mixture parameters, which reduces the computational burden. However, the GM-PHD filter only considers the target's linear motion and linear measurement model. In practical applications, sensors such as radar obtain nonlinear distance and angle information in the measurement equation. Faced with nonlinear models, the filter will have problems with inaccurate estimation of target status or target number.

[0004] Therefore, in order to take into account both real-time performance and tracking accuracy, the GM-PHD filter is considered to be improved to make it suitable for nonlinear environments. In single-target filtering, the analytical approximation of nonlinear Bayesian filters mainly includes extended Kalman filters (EKF) and insensitive Kalman filters (UKF). Therefore, a similar method can be used to extend GM-PHD to compatible nonlinear Gaussian motion models and measurement models. Similar to the single-target case, the EKF-PHD filter linearizes the nonlinear system through Taylor series expansion. The approximate algorithm leads to large errors and can only be applied to differentiable nonlinear models. In addition, the calculation of the Jacobian matrix is ​​lengthy and error-prone. The UKF-PHD filter uses traceless changes to deal with the nonlinear transfer problem of the mean and covariance. At the same time, it does not need to derive the Jacobian matrix and does not ignore high-order terms. Therefore, it has higher calculation accuracy for the statistics of nonlinear distribution. The algorithm is simple and accurate, which is more suitable for radar monitoring nonlinear scenes.

[0005] Holographic staring radar echoes contain a large amount of characteristic information, such as distance, azimuth, amplitude, signal-to-noise ratio, Doppler information, etc. Since some characteristic information between targets and clutter, and between targets and targets is different, one or several characteristic information in the radar echo can be used to distinguish targets from clutter and different targets. The characteristic information can be used to correct the weight of the Gaussian component, thereby greatly reducing the weight of clutter and preventing the merging of different targets. In scenarios with a large number of clutter and a low detection probability, the radar multi-target tracking accuracy can be effectively improved. In addition, in the pruning and merging steps, Doppler information can be used to further eliminate clutter, reduce iterative components, and improve real-time performance.

[0006] In the field of multi-target tracking algorithms based on random finite sets, there are mainly the following problems:

[0007] 1. Few people pay attention to the multi-target tracking method using radar echo information, and characteristic information such as Doppler, amplitude, and signal-to-noise ratio has not been effectively utilized.

[0008] 2. The existing feature-assisted GM-PHD filter is only applicable to Gaussian linear models, and the tracking error is large when tracking nonlinear targets.

[0009] 3. The existing feature-assisted SMC-PHD filter requires a large number of particles and additional clustering operations to extract multi-target state estimates, which has high computational cost and is difficult to implement in engineering.

[0010] 4. Such algorithms are rarely used in actual engineering and lack verification with measured data. Summary of the invention

[0011] The main purpose of the embodiments of the present invention is to provide a nonlinear scene multi-target tracking method, system and storage medium with high tracking accuracy and low cost, which can improve the effectiveness of the method.

[0012] To achieve the above object, an embodiment of the present invention provides a method for tracking multiple targets in a nonlinear scene, comprising the following steps:

[0013] Constructing a nonlinear motion model and a measurement model, specifically including: representing the nonlinear state process and measurement process through a nonlinear model; determining the motion equation of each target according to multi-target tracking in a two-dimensional plane, setting the motion equation of each target as a collaborative turning model, and determining the motion equation of the target; constructing a system state transfer matrix, a process noise distribution matrix, and a measurement equation;

[0014] According to the nonlinear motion model and the measurement model, the multi-target state estimation is extracted by the FAUKF-PHD filter to obtain the state vector of each target and complete the tracking of each target;

[0015] Based on multiple preset scene information, the tracking results of each target are verified.

[0016] In some embodiments, the step of constructing the nonlinear motion model and the measurement model comprises the following steps:

[0017] Represent nonlinear state processes through nonlinear models;

[0018] Considering multiple target tracking in a two-dimensional plane, the motion equation of each target is set as a cooperative turning model.

[0019] In some embodiments, the expression of the nonlinear state process is:

[0020] x k =f k|k-1 (x k-1 ,q k-1 )

[0021] z k =h k (x k , r k )

[0022] Among them, x k and z k They represent the target state and measurement set at time k respectively; f k|k-1 and h k are the nonlinear state transfer function and the measurement function respectively; Gaussian noise q k-1 and r k are zero-mean process noise and measurement noise, respectively, and the corresponding covariance matrices are Q k-1 and R k ;

[0023] The expression of the motion equation of the target is:

[0024] x k =F(ω k-1 )x k-1 +Gω k-1

[0025] Where F represents the system state transfer matrix; G represents the process noise distribution matrix; ω k-1 represents process noise.

[0026] In some embodiments, the expression of the system state transfer matrix F is:

[0027]

[0028] Where T represents the sampling interval;

[0029] The expression of the process noise distribution matrix G is:

[0030]

[0031] The expression of the process noise is:

[0032]

[0033] Among them, v k is a Gaussian white noise with zero mean, x x,k and x y，k Represents the position state of the target at time k.

[0034] In some embodiments, the extraction of multi-target state estimation by using a FAUKF-PHD filter based on the nonlinear motion model and the measurement model to obtain the state vector of each target includes the following steps:

[0035] Construct a Gaussian mixture form of the posterior intensity at time k-1 to obtain a new measurement set;

[0036] Based on the new measurement set, the mean, covariance and weight of the posterior strength are calculated based on the mean, covariance and weight of the predicted strength;

[0037] According to the feature-assisted UKF-PHD filter, the Gaussian component weights are modified using the high-precision Doppler information in the holographic staring radar echo.

[0038] The minor components are pruned according to the Gaussian mixture pruning technique, and similar components are merged;

[0039] According to the final Gaussian mixture expression, the state vectors of all targets are extracted.

[0040] In some embodiments, in the step of constructing a Gaussian mixture form of the posterior intensity at time k-1 to obtain a new measurement set, the expression of the Gaussian mixture form is:

[0041]

[0042] Among them, v k-1 (x) represents the posterior strength at time k-1; J k-1 Represents the total number of given new target components; represent the weight, mean, and covariance of the i-th mixture component representing the nascent intensity, respectively; represents a Gaussian density with mean m and covariance P;

[0043] The expression for the process of calculating the mean, covariance and weight of the posterior strength according to the new measurement set and the mean, covariance and weight of the predicted strength is:

[0044]

[0045] Among them, v k (x) represents the posterior strength at time k; p D,k represents the target detection probability; Z k Represents the information of a certain measurement point at time k;

[0046] The step of correcting the Gaussian component weight using the high-precision Doppler information in the holographic staring radar echo according to the feature-assisted UKF-PHD filter includes the following steps:

[0047] Using the position information in the measurement set, a Gaussian probability density function is used to describe the degree of match between the position information in the measurement set and the position information contained in the predicted Gaussian component;

[0048] Update the weights of the Gaussian components using the corrected likelihood values;

[0049] Substitute the new weights into the posterior strength to complete the feature matching step;

[0050] The step of extracting the state vectors of all targets based on the Gaussian mixture expression finally obtained includes the following steps:

[0051] According to the Gaussian mixture expression of the posterior intensity, the sum of the weights is obtained to estimate the number of targets;

[0052] The component with the largest weight corresponding to the number is selected as the state estimate, or the mean of the Gaussian components with weights greater than a preset threshold is extracted as the state vector of all targets existing at time k.

[0053] In some embodiments, the verification of the tracking results of each target based on the plurality of preset scene information includes the following steps:

[0054] Based on the preset first detection probability and the first clutter number, the tracking performance of the filter is compared from the average OSPA distance and the number of targets, and the real-time performance of the filter is compared from the running time;

[0055] Based on the preset detection probability, the number of clutter is gradually increased to determine how the average OSPA distance of different filters changes with the increase in the number of clutter;

[0056] Based on the preset number of clutter, the detection probability is gradually reduced to determine the change of the average OSPA distance of different filters as the detection probability decreases;

[0057] Based on the preset second detection probability and second clutter number, the target quantity comparison of different filters is determined to complete the verification of the tracking results of each target.

[0058] Another aspect of the embodiment of the present invention further provides a nonlinear scene multi-target tracking system, comprising:

[0059] The first module is used to construct a nonlinear motion model and a measurement model, specifically including: representing the nonlinear state process and measurement process through a nonlinear model; determining the motion equation of each target according to multi-target tracking in a two-dimensional plane, setting it as a collaborative turning model, and determining the motion equation of the target; constructing a system state transfer matrix, a process noise distribution matrix, and a measurement equation;

[0060] The second module is used to extract multi-target state estimation through FAUKF-PHD filter according to the nonlinear motion model and the measurement model, obtain the state vector of each target, and complete the tracking of each target;

[0061] The third module is used to verify the tracking results of each target based on multiple preset scene information.

[0062] Another aspect of an embodiment of the present invention further provides an electronic device, including a processor and a memory;

[0063] The memory is used to store programs;

[0064] The processor executes the program to implement the method described above.

[0065] Another aspect of the embodiments of the present invention further provides a computer-readable storage medium, wherein the storage medium stores a program, and the program is executed by a processor to implement the method described above.

[0066] The embodiment of the present invention also discloses a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device can read the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the above method.

[0067] The embodiments of the present invention include at least the following beneficial effects: the present invention provides a nonlinear scene multi-target tracking method, system and storage medium, the scheme by constructing a nonlinear motion model and a measurement model, specifically including: using a nonlinear model to represent a nonlinear state process and a measurement process; according to the multi-target tracking in a two-dimensional plane, determining the motion equation of each target is set as a collaborative turning model to determine the motion equation of the target; constructing a system state transfer matrix, a process noise distribution matrix, and a measurement equation; according to the nonlinear motion model and the measurement model, extracting the multi-target state estimation through a FAUKF-PHD filter, obtaining the state vector of each target, and completing the tracking of each target; based on multiple preset scene information, verifying the tracking results of each target. The present invention has high tracking accuracy and low cost, and can improve the effectiveness of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 It is a schematic diagram of an implementation environment provided by an embodiment of the present invention;

[0069] Figure 2 is a flow chart of the overall steps provided by an embodiment of the present invention;

[0070] Figure 3 is a real trajectory diagram of a target provided by an embodiment of the present invention;

[0071] Figure 4 is a graph showing the change in potential estimates of different filters over time provided by an embodiment of the present invention;

[0072] Figure 5 is a graph showing the average OSPA distance of different filters provided by an embodiment of the present invention changing over time;

[0073] Figure 6 is a comparison diagram of the running time of different filters provided by an embodiment of the present invention;

[0074] Figure 7 is a graph showing the variation of the average OSPA distance of different filters provided by the embodiment of the present invention as the number of clutter increases;

[0075] Figure 8 is a graph showing the average OSPA distance of different filters provided by an embodiment of the present invention as the detection probability decreases;

[0076] Fig. 9 is a comparison diagram of different filter potential estimates in a low detection probability and high clutter number scenario provided by an embodiment of the present invention;

[0077] Fig.10 It is a diagram showing the processing result of measured data by the FAUKF-PHD filter provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present invention. They are only examples of devices and methods consistent with some aspects of the embodiments of the present invention as detailed in the attached claims.

[0079] It is understood that the terms "first", "second", etc. used in the present invention may be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present invention, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0080] The terms "at least one", "multiple", "each", "any", etc. used in the present invention, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0081] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein are only for the purpose of describing the embodiments of the present invention and are not intended to limit the present invention.

[0082] Before describing the embodiments of the present invention in detail, some related technologies involved in the embodiments of the present invention are first described as follows:

[0083] Holographic staring radar: Holographic staring radar is an advanced radar system. Unlike the mechanical scanning mode of traditional radar systems, holographic staring radar can monitor multiple directions in real time without mechanical movement. Its characteristics are that it uses a wide beam for transmission and a staring array for reception to simultaneously form multiple narrow beams to cover the transmission area, which improves the monitoring range and efficiency of the radar system.

[0084] Multi-target tracking: Multi-target tracking technology is a technology used to simultaneously detect, identify and track multiple targets in a dynamic environment. Its main task is to accurately identify the trajectories of multiple targets from the continuous observation data obtained by the sensor, that is, to meet the actual needs of simultaneously obtaining the number of targets and multiple target information from the measurement set of unknown target sources.

[0085] UKF: UKF (Unscented Kalman Filter) is a filtering algorithm used for state estimation of nonlinear systems. UKF approximates nonlinear functions through unscented transformation, that is, it selects a set of deterministic sampling points to approximate the propagation of the distribution, avoiding the first-order linearization of the nonlinear function, thereby improving the filtering accuracy and robustness.

[0086] PHD filtering: PHD (Probability Hypothesis Density) filtering is a Bayesian filtering method for multi-target tracking, which aims to estimate and track the number and state distribution of targets without explicit data association. PHD filtering describes the distribution state of targets in space through a probability hypothesis density function, and predicts and updates the function through a recursive Bayesian update formula.

[0087] GM-PHD filter: GM-PHD (Gaussian Mixture PHD) filter is an advanced filtering algorithm for multi-target tracking. It combines the advantages of PHD filtering and Gaussian mixture model, describes the dynamic distribution state of the target through probability hypothesis density function, and uses Gaussian mixture model to approximate and update this density function, which greatly simplifies the computational complexity.

[0088] SMC-PHD filter: SMC-PHD (Sequential Monte Carlo PHD) filter is a multi-target tracking algorithm based on particle filtering. It combines the theoretical framework of PHD filtering and the numerical computing power of the sequential Monte Carlo method, and uses a group of weighted random particles to approximate and update the probability hypothesis density function, aiming to solve the problem of target quantity and state estimation in a dynamic environment.

[0089] The nonlinear scene multi-target tracking method, system and storage medium provided by the embodiment of the present invention relate to the field of computer technology. The nonlinear scene multi-target tracking method provided by the embodiment of the present invention can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements the nonlinear scene multi-target tracking method, etc., but is not limited to the above forms.

[0090] The present invention can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present invention can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0091] like Figure 1 FIG. 1 is a schematic diagram of an implementation environment provided by an embodiment of the present invention. Figure 1 The implementation environment includes at least one terminal 102 and a server 101. The terminal 102 and the server 101 can be connected to a network wirelessly or wired to complete data transmission and exchange.

[0092] Server 101 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), as well as big data and artificial intelligence platforms.

[0093] In addition, the server 101 can also be a node server in the blockchain network. Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithm.

[0094] The terminal 102 may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc. The terminal 102 may also be a vehicle-mounted terminal of various device types described above, but is not limited thereto. The terminal 102 and the server 101 may be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment of the present invention.

[0095] Based on the example Figure 1 In the implementation environment shown, an embodiment of the present invention provides a nonlinear scene multi-target tracking method. The following is an example of applying the nonlinear scene multi-target tracking method in the server 101. It can be understood that the method can also be applied to the terminal 102.

[0096] Reference Figure 2 , Figure 2 The flowchart of the method for tracking multiple targets in a nonlinear scene applied to a server provided by an embodiment of the present invention, the execution subject of the method can be any of the aforementioned computer devices (including a server or a terminal). Figure 2 , the method may include the following steps:

[0097] Constructing a nonlinear motion model and a measurement model, specifically including: representing the nonlinear state process and measurement process through a nonlinear model; determining the motion equation of each target according to multi-target tracking in a two-dimensional plane, setting the motion equation of each target as a collaborative turning model, and determining the motion equation of the target; constructing a system state transfer matrix, a process noise distribution matrix, and a measurement equation;

[0098] According to the nonlinear motion model and the measurement model, the multi-target state estimation is extracted by the FAUKF-PHD filter to obtain the state vector of each target and complete the tracking of each target;

[0099] Based on multiple preset scene information, the tracking results of each target are verified.

[0100] In some embodiments, the step of constructing the nonlinear motion model and the measurement model comprises the following steps:

[0101] Represent nonlinear state processes through nonlinear models;

[0102] Considering multiple target tracking in a two-dimensional plane, the motion equation of each target is set as a cooperative turning model.

[0103] In some embodiments, the expression of the nonlinear state process is:

[0104] x k =f k|k-1 (x k-1 ,q k-1 )

[0105] z k =h k (x k , r k )

[0106] Among them, x k and z k They represent the target state and measurement set at time k respectively; f k|k-1 and h k are the nonlinear state transfer function and the measurement function respectively; Gaussian noise q k-1 and r k are zero-mean process noise and measurement noise, respectively, and the corresponding covariance matrices are Q k-1 and R k ;

[0107] The expression of the motion equation of the target is:

[0108] x k =F(ω k-1 )x k-1 +Gω k-1

[0109] Where F represents the system state transfer matrix; G represents the process noise distribution matrix; ω k-1 represents process noise.

[0110] In some embodiments, the expression of the system state transfer matrix F is:

[0111]

[0112] Where T represents the sampling interval;

[0113] The expression of the process noise distribution matrix G is:

[0114]

[0115] The expression of the process noise is:

[0116]

[0117] Among them, v k is a Gaussian white noise with zero mean, x x,k and x y，k Represents the position state of the target at time k.

[0118] In some embodiments, the extraction of multi-target state estimation by using a FAUKF-PHD filter based on the nonlinear motion model and the measurement model to obtain the state vector of each target includes the following steps:

[0119] Construct a Gaussian mixture form of the posterior intensity at time k-1 to obtain a new measurement set;

[0120] Based on the new measurement set, the mean, covariance and weight of the posterior strength are calculated based on the mean, covariance and weight of the predicted strength;

[0121] According to the feature-assisted UKF-PHD filter, the Gaussian component weights are modified using the high-precision Doppler information in the holographic staring radar echo.

[0122] The minor components are pruned according to the Gaussian mixture pruning technique, and similar components are merged;

[0123] According to the final Gaussian mixture expression, the state vectors of all targets are extracted.

[0124] In some embodiments, in the step of constructing a Gaussian mixture form of the posterior intensity at time k-1 to obtain a new measurement set, the expression of the Gaussian mixture form is:

[0125]

[0126] Among them, v k-1 (x) represents the posterior strength at time k-1; J k-1 Represents the total number of given new target components; represent the weight, mean, and covariance of the i-th mixture component representing the nascent intensity, respectively; represents a Gaussian density with mean m and covariance P;

[0127] The expression for the process of calculating the mean, covariance and weight of the posterior strength according to the new measurement set and the mean, covariance and weight of the predicted strength is:

[0128]

[0129] Among them, v k (x) represents the posterior strength at time k; p D,k represents the target detection probability; Z k Represents the information of a certain measurement point at time k;

[0130] The step of correcting the Gaussian component weight using the high-precision Doppler information in the holographic staring radar echo according to the feature-assisted UKF-PHD filter includes the following steps:

[0131] Using the position information in the measurement set, a Gaussian probability density function is used to describe the degree of match between the position information in the measurement set and the position information contained in the predicted Gaussian component;

[0132] Update the weights of the Gaussian components using the corrected likelihood values;

[0133] Substitute the new weights into the posterior strength to complete the feature matching step;

[0134] The step of extracting the state vectors of all targets based on the Gaussian mixture expression finally obtained includes the following steps:

[0135] According to the Gaussian mixture expression of the posterior intensity, the sum of the weights is obtained to estimate the number of targets;

[0136] The component with the largest weight corresponding to the number is selected as the state estimate, or the mean of the Gaussian components with weights greater than a preset threshold is extracted as the state vector of all targets existing at time k.

[0137] In some embodiments, the verification of the tracking results of each target based on the plurality of preset scene information includes the following steps:

[0138] Based on the preset first detection probability and the first clutter number, the tracking performance of the filter is compared from the average OSPA distance and the number of targets, and the real-time performance of the filter is compared from the running time;

[0139] Based on the preset detection probability, the number of clutter is gradually increased to determine how the average OSPA distance of different filters changes with the increase in the number of clutter;

[0140] Based on the preset number of clutter, the detection probability is gradually reduced to determine the change of the average OSPA distance of different filters as the detection probability decreases;

[0141] Based on the preset second detection probability and second clutter number, the target quantity comparison of different filters is determined to complete the verification of the tracking results of each target.

[0142] The implementation process of the present invention in a specific scenario is described in detail below in conjunction with the accompanying drawings of the specification:

[0143] In view of the problems existing in the prior art, the object of the present invention is to provide a radar multi-target tracking method based on a feature-assisted UKF-PHD filter, and the present invention has the following characteristics:

[0144] 1. The holographic staring radar can obtain high Doppler resolution by accumulating data over a long period of time. The Doppler information in the radar echo can be combined to distinguish between targets and clutter, and between targets, thereby improving tracking accuracy.

[0145] 2. The UKF-PHD filter is selected as the basic filter for improvement, which not only expands the GM-PHD filter to nonlinear scenarios, but also takes into account the advantage of low computational cost of the GM-PHD filter.

[0146] 3. The effectiveness of the algorithm is verified by the measured data collected by the holographic staring radar.

[0147] The specific operation steps of the present invention are as follows:

[0148] Step 1: Modeling

[0149] The nonlinear state process and measurement process can be represented by the following nonlinear model, namely:

[0150] x k =f k|k-1 (x k-1 ,q k-1 ) (1)

[0151] z k =h k (x k , r k ) (2)

[0152] Among them, x k and z k They represent the target state and measurement set at time k respectively. k|k-1 and h k are the nonlinear state transfer function and the measurement function respectively; Gaussian noise q k-1 and r k are zero-mean process noise and measurement noise, respectively, and the corresponding covariance matrices are Q k-1 and R k Since the state function and the measurement function are nonlinear functions, the posterior strength of the multi-target state predicted and updated by the PHD filter is the weighted sum of different non-Gaussian functions. The UKF can be used to approximate each non-Gaussian function with a single Gaussian. The specific operation is to use an insensitive transformation to calculate the components of the Gaussian mixture approximation of the posterior strength.

[0153] Considering multi-target tracking in a two-dimensional plane, the motion equation of each target is set as a coordinated turn (CT) model, then the motion equation of the target is:

[0154] x k =F(ω k-1 )x k-1 +Gω k-1 (3)

[0155] Where, process noise ω k-1 are mutually independent zero-mean Gaussian white noises, and the system state transfer matrix is:

[0156]

[0157] The corresponding process noise distribution matrix is:

[0158]

[0159] The measurement equation of the target is:

[0160]

[0161] Among them, v k is a Gaussian white noise with zero mean, x x，k and x y，k Represents the position state of the target at time k.

[0162] Step 2: FAUKF-PHD filter

[0163] Step 1: Assume that the posterior strength v at time k-1 k-1 It is in the form of a Gaussian mixture:

[0164]

[0165] The prediction step is based on v k-1 The mean, covariance and weights of k|k-1 The process of calculating the mean, covariance and weight expression of . Note that the prediction intensity v at time k is also in the form of Gaussian mixture k|k-1 As shown below:

[0166]

[0167] Among them, v γ,k (·) is the strength of the new target at time k, v S,k|k-1 (·) is the survival target strength at time k. According to the given Gaussian component set Record the mean Covariance Using insensitive transformation, a series of sigma points and their weights are generated according to the mean and covariance, recorded as The N value is twice the covariance dimension. In addition, there are some parameters that control the performance of the UKF filter. The a parameter determines the confidence level of the state estimate in the UKF, and the β parameter is used to measure the uncertainty between the Gaussian distribution of the state and the Gaussian distribution of the measurement. The parameter affects the choice of sigma points. For n = 0, 1, ..., N, the split Then the predicted strength v at time k is k|k-1 The relevant variable expressions are as follows:

[0168]

[0169] Step 2: Update

[0170] Once the new measurement set Z is obtained k , the update step provides the k|k-1 The mean, covariance and weights of k The steps of the mean, covariance and weight expressions. Let the posterior strength v at time k be the same Gaussian mixture form k As shown below:

[0171]

[0172] Then the posterior strength v at time k is k The relevant variable expressions are as follows:

[0173]

[0174] Among them, κ k is the intensity function of the Poisson distributed clutter RFS at time k.

[0175] Step 3: Feature Assistance

[0176] The feature-assisted UKF-PHD filter proposed in the present invention uses the high-precision Doppler information in the holographic staring radar echo to correct the Gaussian component weight, increase the weight of the target component, and reduce the weight of the clutter component, thereby achieving the distinction between target and clutter, target and target, and improving the radar multi-target tracking accuracy. After the weight is corrected, more non-correlated Gaussian components can be filtered out in the pruning and merging steps, thereby improving the radar multi-target tracking efficiency.

[0177] According to the expression of Gaussian component weight in the update step, the variable that affects the weight is This formula only uses the position information in the measurement, and adopts the Gaussian probability density function to describe the degree of match between the position information of the measurement set and the position information contained in the predicted Gaussian component. In order to update the Gaussian component weights using the characteristic information, a Gaussian probability density function is added to describe the degree of match between the Doppler information of the measurement set and the Doppler information contained in the predicted Gaussian component. The expression for the Doppler frequency characteristic likelihood is as follows:

[0178]

[0179] Among them, z D represents the measured Doppler frequency. In the present invention, f d，k represents the Doppler frequency of the target at time k, It represents the overall error between the Doppler information in the measurement set and the Doppler information of the predicted Gaussian component. The greater the difference in Doppler frequency between different targets, the better the tracking performance. The greater the difference in Doppler frequency between the target and clutter, the better the tracking performance. The smaller the change in Doppler frequency of the same target between two adjacent frames, the better the tracking performance. The expression after adding feature information is:

[0180]

[0181] Among them, z all Represents a measurement set containing location information and Doppler information. Then the weight of the Gaussian component is updated with the modified likelihood value. The modified weight expression is:

[0182]

[0183] Substitute the new weight into the posterior strength v k That is, the feature matching step is completed. At this time, in addition to the position information of all Gaussian components, the Gaussian component also incorporates the Doppler information contained in all Gaussian components. Therefore, the feature auxiliary step can be placed before the update step.

[0184] Step 4: Pruning and merging

[0185] As the processing time increases, the number of Gaussian mixture components in the FAUKF-PHD filter will continue to increase. In order to effectively control the growing number of Gaussian components, it is necessary to use Gaussian mixture pruning techniques such as pruning minor components and merging similar components. Set the pruning threshold T to discard components with weights below the preset threshold; set the maximum allowable number of Gaussian terms J max , when the number of components exceeds this value, only the components with the corresponding number of maximum weights are retained; set the merging threshold U to merge some Gaussian components that are close enough.

[0186] Step 5: Status extraction

[0187] The FAUKF-PHD filter is easy to extract multi-target state estimates based on the posterior strength v k The Gaussian mixture expression can first obtain the sum of the weights to estimate the number of targets, and then select the component with the largest weight corresponding to the number as the state estimate, or extract the mean of the Gaussian components with weights greater than a certain threshold as the state vector of all targets existing at time k.

[0188] Step 3: Results display of the present invention

[0189] The simulation experiment considers multi-target tracking in a two-dimensional plane, and sets the monitoring area to [-2000 2000]×

[02000] (unit: m). The number of targets is unknown and time-varying. In order to verify the effectiveness of the proposed FAUKF-PHD filter, different scenarios are designed and the results are compared with the standard GM-PHD, SMC-PHD and UKF-PHD. Figure 3 is the real trajectory of the target, with common scenarios such as target birth, target extinction, and trajectory crossing. Under the same target motion trajectory, 100 Monte Carlo experiments are run, each experiment is independent of each other, and the performance of each filter is compared. Figure 3 In the figure, True track represents the true track; Track start represents the starting point; Track end represents the ending point.

[0190] Scenario 1: When the detection probability p D,k =0.98, clutter number λ=30, the tracking performance of the filter is compared from the average OSPA distance and the number of targets, and the real-time performance of the filter is compared from the running time. The comparison of the number of targets, average OSPA distance, and running time of different filters is shown in the figure below. Figure 4 , Figure 5 , Figure 6 shown.

[0191] in, Figure 4 The horizontal and vertical axes of are time and target quantity respectively; Figure 5 The horizontal and vertical axes of are time and OSPA distance respectively; Figure 6 The horizontal and vertical axes are the number of Monte Carlo experiments and the running time, respectively.

[0192] Depend on Figure 4 , Figure 5 It can be seen that under the nonlinear motion model and measurement model, the tracking performance of the GM-PHD filter is very poor and the number of targets cannot be accurately estimated. The number of targets estimated by the proposed FAUKF-PHD filter is closest to the actual number of targets and is the most stable in the tracking estimation process, and the average OSPA distance is also kept to the minimum. Figure 6It can be seen that the computational cost of the SMC-PHD filter is high, and the running time of the proposed algorithm is close to that of the GM-PHD filter, with high computational efficiency. Therefore, in scenarios with low clutter rate and high detection probability, the FAUKF-PHD filter improves the radar multi-target tracking performance and potential estimation performance without causing a sharp increase in time cost.

[0193] Scenario 2: When the detection probability p D,k =0.98, when the number of clutter λ increases, the average OSPA distance of different filters changes with the increase of the number of clutter as shown in the figure below: Figure 7 shown. Figure 7 The horizontal and vertical coordinates of are the clutter number and OSPA distance; Figure 7 It can be seen that the average OSPA distance of each filter increases with the increase of the number of clutter, while the average OSPA distance of the FAUKF-PHD filter remains almost unchanged, is less affected by the increase of the number of clutter, and has the best tracking performance. Therefore, the FAUKF-PHD filter can adapt to radar multi-target tracking in dense clutter environments.

[0194] Scenario 3: When the number of clutter λ = 30, the detection probability p D,k As the detection probability decreases, the average OSPA distance of different filters changes as the detection probability decreases. Figure 8 shown. Figure 8 The horizontal and vertical axes are the detection probability and OSPA distance; Figure 8 It can be seen that as the detection probability continues to decrease, the average OSPA distance of each filter continues to increase, but the average OSPA distance of the FAUKF-PHD filter is the lowest. Therefore, the FAUKF-PHD filter can adapt to radar multi-target tracking in low detection probability environments.

[0195] Scenario 4: When the detection probability p D,k =0.85, clutter number λ = 150, the target number comparison of different filters is as follows Fig. 9 shown. Fig. 9 The horizontal and vertical axes of Fig. 9 It can be seen that when the detection probability is low and the number of clutter is high, the number of targets estimated by the FAUKF-PHD filter is close to the actual number of targets, while the standard UKF-PHD filter has obvious missed detections. The potential estimation performance of the improved filter is better than that of the standard UKF-PHD filter.

[0196] Measured scenario: The proposed method was verified using multi-target data collected by a holographic staring radar system. The radar was installed at a high point along the coast and used to track two drones flying in eccentric circles on the sea surface. The collected real-time data was processed using the FAUKF-PHD filter. The results are shown in Figure 2. Fig.10 shown. Fig.10 The horizontal and vertical coordinates of are angles and radial distances; the legend Measurements is the measured value, and Estimates is the estimated value. It can be seen that despite the presence of ocean clutter, the filtering results are still evenly distributed around the target measurement point, demonstrating the effectiveness of the proposed algorithm.

[0197] In summary, the present invention adopts a feature-assisted method to optimize the UKF-PHD filter, which is suitable for nonlinear scenes and has low computational cost, to achieve multi-target tracking in clutter-intensive scenes.

[0198] Compared with the prior art, the present invention has the following characteristics:

[0199] 1. Improved UKF-PHD filter: Considering the high Doppler resolution of holographic staring radar, a UKF-PHD filter based on Doppler feature assistance is proposed to improve the tracking accuracy in dense clutter scenes.

[0200] 2. Measured data verification: Use the algorithm to process the nonlinear multi-target measured data collected by the holographic staring radar to verify that the algorithm takes into account both tracking efficiency and tracking accuracy.

[0201] Compared with the prior art, the present invention has the following advantages:

[0202] Compared with the optimized GM-PHD filter, the present invention selects the UKF-PHD filter for improvement. While retaining the small computational complexity of the GM-PHD filter, it extends it to nonlinear scenarios, which is more in line with engineering reality. The high-precision Doppler information in the holographic staring radar echo is used to correct the Gaussian component weights, and the accuracy of the weights can be greatly improved by using only one feature information. The measured data are used to verify the effectiveness of the algorithm in multi-target tracking under dense clutter and low detection probability scenarios.

[0203] Another aspect of the embodiment of the present invention further provides a nonlinear scene multi-target tracking system, comprising:

[0204] The first module is used to construct a nonlinear motion model and a measurement model, specifically including: representing the nonlinear state process and measurement process through a nonlinear model; determining the motion equation of each target according to multi-target tracking in a two-dimensional plane, setting it as a collaborative turning model, and determining the motion equation of the target; constructing a system state transfer matrix, a process noise distribution matrix, and a measurement equation;

[0205] The second module is used to extract multi-target state estimation through FAUKF-PHD filter according to the nonlinear motion model and the measurement model, obtain the state vector of each target, and complete the tracking of each target;

[0206] The third module is used to verify the tracking results of each target based on multiple preset scene information.

[0207] It can be understood that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0208] The embodiment of the present invention further provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned nonlinear scene multi-target tracking method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0209] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0210] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the nonlinear scene multi-target tracking method is implemented.

[0211] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0212] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0213] It should be noted that in various specific embodiments of the present invention, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present invention needs to obtain the user's sensitive personal information, it will obtain the user's separate permission or consent through a pop-up window or jump to a confirmation page, and after clearly obtaining the user's separate permission or consent, it will obtain the necessary user-related data for the normal operation of the embodiment of the present invention.

[0214] The embodiments described in the embodiments of the present invention are intended to more clearly illustrate the technical solutions of the embodiments of the present invention, and do not constitute a limitation on the technical solutions provided by the embodiments of the present invention. Those skilled in the art can appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present invention are also applicable to similar technical problems.

[0215] Those skilled in the art will appreciate that the technical solutions shown in the figures do not limit the embodiments of the present invention and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0216] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0217] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0218] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0219] It should be understood that in the present invention, "at least one (item)" refers to one or more, and "plurality" refers to two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can represent: only A exists, only B exists, and A and B exist at the same time, 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 items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0220] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as 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.

[0221] The units described above 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.

[0222] In addition, each functional unit in each embodiment of the present invention 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. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0223] If the integrated unit 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 invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store programs.

[0224] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the embodiments of the present invention is not limited thereby. Any modification, equivalent substitution and improvement made by those skilled in the art without departing from the scope and essence of the embodiments of the present invention shall be within the scope of the rights of the embodiments of the present invention.

Claims

1. A method for tracking multiple targets in a nonlinear scene, characterized in that: The following steps are involved: Constructing a nonlinear motion model and a measurement model, specifically including: representing the nonlinear state process and measurement process through a nonlinear model; determining the motion equation of each target according to multi-target tracking in a two-dimensional plane, setting the motion equation of each target as a collaborative turning model, and determining the motion equation of the target; constructing a system state transfer matrix, a process noise distribution matrix, and a measurement equation; According to the nonlinear motion model and the measurement model, the multi-target state estimation is extracted by the FAUKF-PHD filter to obtain the state vector of each target and complete the tracking of each target; Based on multiple preset scene information, the tracking results of each target are verified.

2. A nonlinear scene multi-target tracking method according to claim 1, characterized in that: The construction of the nonlinear motion model and the measurement model comprises the following steps: Represent nonlinear state processes through nonlinear models; Considering multiple target tracking in a two-dimensional plane, the motion equation of each target is set as a cooperative turning model.

3. A nonlinear scene multi-target tracking method according to claim 2, characterized in that: The expression of the nonlinear state process is: x k =f k|k-1 (x k-1 ,q k-1 ) z k =h k (x k ,r k ) Among them, x k and z k They represent the target state and measurement set at time k respectively; f k|k-1 and h k are the nonlinear state transfer function and the measurement function respectively; Gaussian noise q k-1 and r k are zero-mean process noise and measurement noise, respectively, and the corresponding covariance matrices are Q k-1 and R k ; The expression of the motion equation of the target is: x k =F(ω k-1 )x k-1 +Gω k-1 Where F represents the system state transfer matrix; G represents the process noise distribution matrix; ω k-1 represents process noise.

4. The method for tracking multiple targets in a nonlinear scene according to claim 3, characterized in that: The expression of the system state transfer matrix F is: Where T represents the sampling interval; The expression of the process noise distribution matrix G is: The expression of the process noise is: Among them, v k is a Gaussian white noise with zero mean, x x，k and x y，k Represents the position state of the target at time k.

5. The method for tracking multiple targets in a nonlinear scene according to claim 1, characterized in that: The method of extracting multi-target state estimation by using a FAUKF-PHD filter according to the nonlinear motion model and the measurement model to obtain a state vector of each target includes the following steps: Construct a Gaussian mixture form of the posterior intensity at time k-1 to obtain a new measurement set; Based on the new measurement set, the mean, covariance and weight of the posterior strength are calculated based on the mean, covariance and weight of the predicted strength; According to the feature-assisted UKF-PHD filter, the Gaussian component weights are modified using the high-precision Doppler information in the holographic staring radar echo. The minor components are pruned according to the Gaussian mixture pruning technique, and similar components are merged; According to the final Gaussian mixture expression, the state vectors of all targets are extracted.

6. A nonlinear scene multi-target tracking method according to claim 5, characterized in that: In the step of constructing the Gaussian mixture form of the posterior intensity at time k-1 to obtain a new measurement set, the expression of the Gaussian mixture form is: Among them, v k-1 (x) represents the posterior strength at time k-1; J k-1 Represents the total number of given new target components; represent the weight, mean, and covariance of the i-th mixture component representing the nascent intensity, respectively; represents a Gaussian density with mean m and covariance P; The expression for calculating the mean, covariance and weight of the posterior strength according to the new measurement set and the mean, covariance and weight of the predicted strength is: Among them, v k (x) represents the posterior strength at time k; p D，k represents the target detection probability; Z k Represents the information of a certain measurement point at time k; The step of correcting the Gaussian component weight using the high-precision Doppler information in the holographic staring radar echo according to the feature-assisted UKF-PHD filter includes the following steps: Using the position information in the measurement set, a Gaussian probability density function is used to describe the degree of match between the position information in the measurement set and the position information contained in the predicted Gaussian component; Update the weights of the Gaussian components using the corrected likelihood values; Substitute the new weights into the posterior strength to complete the feature matching step; The step of extracting the state vectors of all targets based on the Gaussian mixture expression finally obtained includes the following steps: According to the Gaussian mixture expression of the posterior intensity, the sum of the weights is obtained to estimate the number of targets; The component with the largest weight corresponding to the number is selected as the state estimate, or the mean of the Gaussian components with weights greater than a preset threshold is extracted as the state vector of all targets existing at time k.

7. The method for tracking multiple targets in a nonlinear scene according to claim 1, characterized in that: The verification of the tracking results of each target based on the plurality of preset scene information includes the following steps: Based on the preset first detection probability and the first clutter number, the tracking performance of the filter is compared from the average OSPA distance and the number of targets, and the real-time performance of the filter is compared from the running time; Based on the preset detection probability, the number of clutter is gradually increased to determine how the average OSPA distance of different filters changes with the increase in the number of clutter; Based on the preset number of clutter, the detection probability is gradually reduced to determine the change of the average OSPA distance of different filters as the detection probability decreases; Based on the preset second detection probability and second clutter number, the target quantity comparison of different filters is determined to complete the verification of the tracking results of each target.

8. A nonlinear scene multi-target tracking system, characterized in that: include: The first module is used to construct a nonlinear motion model and a measurement model, specifically including: representing the nonlinear state process and measurement process through a nonlinear model; determining the motion equation of each target according to multi-target tracking in a two-dimensional plane, setting it as a collaborative turning model, and determining the motion equation of the target; constructing a system state transfer matrix, a process noise distribution matrix, and a measurement equation; The second module is used to extract multi-target state estimation through FAUKF-PHD filter according to the nonlinear motion model and the measurement model, obtain the state vector of each target, and complete the tracking of each target; The third module is used to verify the tracking results of each target based on multiple preset scene information.

9. An electronic device, characterized in that: including a processor and a memory; The memory is used to store programs; The processor executes the program to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The storage medium stores a program, and the program is executed by a processor to implement the method according to any one of claims 1 to 7.

Images