阅读说明：本技术 一种基于特征的目标跟踪优化方法、装置和存储介质 (Target tracking optimization method and device based on features and storage medium ) 是由 陶烨 屈操 闫红宇 于 2021-09-14 设计创作，主要内容包括：本发明提供一种基于特征的目标跟踪优化方法,包括以下步骤：步骤S10,对得到的目标点迹信息进行处理,分别形成静目标轨迹和动目标轨迹；进行静目标数据关联,得到与静目标轨迹关联的静目标；进行动目标数据关联,得到与动目标轨迹关联的动目标；步骤S20,基于与动目标轨迹关联的动目标利用IMM算法进行动目标跟踪。本发明还提出了一种基于特征的目标跟踪优化装置,包括：存储器,存储有计算机程序；处理器,所述处理器包含数据处理单元,所述数据处理单元用于运行所述计算机程序,所述计算机程序运行时执行如上文所述的方法的步骤。本发明能有效的实时跟踪目标,准确度高,在工程上更具适用性。(The invention provides a target tracking optimization method based on characteristics, which comprises the following steps: step S10, processing the obtained target trace information to respectively form a static target trace and a moving target trace; performing static target data association to obtain a static target associated with a static target track; performing moving target data association to obtain a moving target associated with a moving target track; and step S20, performing moving target tracking by utilizing an IMM algorithm based on the moving target related to the moving target track. The invention also provides a target tracking optimization device based on the characteristics, which comprises the following steps: a memory storing a computer program; a processor comprising a data processing unit for executing the computer program, which computer program when executed performs the steps of the method as described above. The invention can effectively track the target in real time, has high accuracy and is more applicable to engineering.)

1. A target tracking optimization method based on characteristics is characterized by comprising the following steps:

step S10, processing the obtained target trace information to respectively form a static target trace and a moving target trace; performing static target data association to obtain a static target associated with a static target track; performing moving target data association to obtain a moving target associated with a moving target track;

and step S20, performing moving target tracking by utilizing an IMM algorithm based on the moving target related to the moving target track.

2. The feature-based object tracking optimization method of claim 1,

the step S10 specifically includes:

step S11, clustering the targets;

step S12, separating the moving and static targets according to the clustered speed information and characteristic information;

step S13, forming a static target track after the separated static target is initialized;

step S14, performing static target data association according to the static target association parameters and the association conditions to obtain a static target associated with the static target track;

step S15, the separated moving target forms a moving target track after being initialized;

and step S16, performing moving target data association according to the moving target association parameters and the association conditions to obtain a moving target associated with the moving target track.

3. The feature-based object tracking optimization method of claim 1,

the IMM algorithm includes a plurality of object motion models, and the step S20 specifically includes:

step S21, inputting the posterior probability, Markov transition probability and covariance matrix of each model obtained by calculation in the previous period as input, and the information of the moving target related to the moving target track into each model;

the posterior probability of each model refers to the probability distributed by each model calculated according to the measurement value in the previous period; the Markov transition probability is an empirical value matrix initially determined according to the conversion possibility among the models;

step S22, carrying out model information interaction: comprehensively calculating according to the posterior probability of each model and the Markov transition probability to obtain a mixed probability; initializing information of a moving target related to a moving target track and a covariance matrix through mixed probability to obtain interactive target information and an interactive covariance matrix;

step S23, interactive target information and an interactive covariance matrix of each model are used as filter input, and the filter of each model is input for parallel filtering;

step S24, obtaining outputs of the respective models in the current period: an estimated value and a covariance matrix;

step S25, model probability updating: calculating the filtering residual error of each model at the current moment by adopting a Bayesian hypothesis test method, and calculating the distribution probability of each model according to the estimated value and the measured value;

step S26, state result interaction: and performing weighted fusion on the estimation value and the covariance matrix of each model according to the distribution probability to obtain a final estimation result of the current period.

4. The feature-based object tracking optimization method of claim 3,

the target motion model at least comprises one of a uniform velocity model and a uniform acceleration model, and one of a Singer model and a current statistical model.

5. A feature-based object tracking optimization apparatus, comprising:

a memory storing a computer program;

processor comprising a data processing unit for executing the computer program, which computer program when executed performs the steps of the method according to any one of claims 1 to 4.

6. A storage medium characterized in that,

the storage medium has stored therein a computer program configured to perform the steps of the method of any one of claims 1 to 4 when executed.

Technical Field

The invention belongs to the field of target tracking of security radar, and particularly relates to a target tracking optimization method based on characteristics in a radar data processing system.

Background

At present, with the development and maturity of millimeter wave radar technology, the application of millimeter wave radar in security systems is more and more extensive. The millimeter wave radar can detect the distance, the radial velocity and the angle of a plurality of targets, compares the camera and does not receive the light influence, and the interference killing feature is strong, can work in all weather for 24 hours. In a security system, a millimeter wave radar needs to track and predict a plurality of targets in an area and simultaneously display the motion tracks of the targets, so that dangerous targets are screened out and monitored and alarmed, and therefore, the maneuvering target tracking is a key link in the radar data processing process, can effectively inhibit random errors introduced in the measuring process, stably estimates and reasonably predicts the motion tracks of the targets, forms stable target tracks and realizes high-precision real-time tracking of the targets. In the process of tracking the moving target, if a tracking algorithm of a single motion model is adopted, the models are set a priori, so that the moving track of the target cannot be well matched, the estimation accuracy of the filter is reduced, the filter is diverged, and even the situations of track fracture and target loss occur.

Disclosure of Invention

The invention aims to overcome the defects in the prior art, and provides a feature-based target tracking optimization method, a feature-based target tracking optimization device and a storage medium, wherein a moving target and a static target are separately processed by utilizing the feature that the static target in a security radar scene environment is basically unchanged, so that the defects of the existing target tracking method are overcome, the target can be effectively tracked in real time, the accuracy is high, and the method has higher applicability in engineering. In order to achieve the technical purpose, the embodiment of the invention adopts the technical scheme that:

in a first aspect, an embodiment of the present invention provides a feature-based target tracking optimization method, including the following steps:

step S10, processing the obtained target trace information to respectively form a static target trace and a moving target trace; performing static target data association to obtain a static target associated with a static target track; performing moving target data association to obtain a moving target associated with a moving target track;

and step S20, performing moving target tracking by utilizing an IMM algorithm based on the moving target related to the moving target track.

Further, the step S10 specifically includes:

step S11, clustering the targets;

step S12, separating the moving and static targets according to the clustered speed information and characteristic information;

step S13, forming a static target track after the separated static target is initialized;

step S14, performing static target data association according to the static target association parameters and the association conditions to obtain a static target associated with the static target track;

step S15, the separated moving target forms a moving target track after being initialized;

and step S16, performing moving target data association according to the moving target association parameters and the association conditions to obtain a moving target associated with the moving target track.

Further, the IMM algorithm includes a plurality of object motion models, and the step S20 specifically includes:

step S21, inputting the posterior probability, Markov transition probability and covariance matrix of each model obtained by calculation in the previous period as input, and the information of the moving target related to the moving target track into each model;

the posterior probability of each model refers to the probability distributed by each model calculated according to the measurement value in the previous period; the Markov transition probability is an empirical value matrix initially determined according to the conversion possibility among the models;

step S22, carrying out model information interaction: comprehensively calculating according to the posterior probability of each model and the Markov transition probability to obtain a mixed probability; initializing information of a moving target related to a moving target track and a covariance matrix through mixed probability to obtain interactive target information and an interactive covariance matrix;

step S23, interactive target information and an interactive covariance matrix of each model are used as filter input, and the filter of each model is input for parallel filtering;

step S24, obtaining outputs of the respective models in the current period: an estimated value and a covariance matrix;

step S25, model probability updating: calculating the filtering residual error of each model at the current moment by adopting a Bayesian hypothesis test method, and calculating the distribution probability of each model according to the estimated value and the measured value;

step S26, state result interaction: and performing weighted fusion on the estimation value and the covariance matrix of each model according to the distribution probability to obtain a final estimation result of the current period.

Preferably, the object motion model at least comprises one of a uniform velocity model and a uniform acceleration model, and one of a Singer model and a current statistical model.

In a second aspect, an embodiment of the present invention further provides a feature-based target tracking optimization apparatus, including:

a memory storing a computer program;

a processor comprising a data processing unit for executing the computer program, which computer program when executed performs the steps of the method as described above.

In a third aspect, an embodiment of the present invention further provides a storage medium, in which a computer program is stored, where the computer program is configured to execute the steps of the method described above when running.

The technical scheme provided by the embodiment of the invention has the following beneficial effects:

1) only millimeter wave radar is used for detecting data, combination with other sensors is not needed, cost and calculation complexity are reduced, influences of factors such as weather are avoided, and usability is high.

2) The moving target and the static target are separated, so that cross correlation is avoided, and the accuracy of target tracking is improved; the method ensures that the moving target can use a complex self-adaptive multi-model interactive tracking algorithm, thereby taking high precision and maneuverability into consideration and meeting the engineering requirements.

3) The moving target tracking uses a self-adaptive multi-model interactive tracking algorithm to be self-adaptive to various scenes, the problem of model divergence caused by a single model under different conditions is solved, and the responsiveness and the stability are improved.

Drawings

Fig. 1 is a general flowchart of a target tracking optimization method in an embodiment of the present invention.

FIG. 2 is a flowchart of an adaptive multi-model interaction tracking algorithm in an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

In the embodiment of the invention, the security radar system mainly comprises a signal transmitting unit, a signal receiving unit, a signal processing unit and a data processing unit; the signal transmitting unit mainly radiates detection signals outwards, and the detection signals are generally sawtooth waves of linear frequency modulation; the signal receiving unit is mainly used for receiving echo signals reflected by a target through a receiving antenna; the signal processing unit processes the received echo signals to obtain the information of the angle, distance and speed (mainly radial speed in a radar system) of a target, namely target trace information; the data processing unit can utilize the target trace information obtained by the signal processing unit to perform algorithms such as clustering and tracking, and the like to process the target trace information, so that the functional requirements required by different radars are met; the operations can be completed in a fixed period, and the operations can be repeated in a continuous cycle in the running process of the radar; it should be noted that the signal processing unit and the data processing unit may be implemented on the same microprocessor, or may be implemented on two microprocessors respectively;

in the embodiment of the invention, a millimeter wave radar in a security radar system needs to continuously track and predict a plurality of targets in an area, and simultaneously displays the motion trail of the targets, so as to screen out dangerous targets, monitor and alarm the dangerous targets; therefore, in order to obtain the continuous change process of the same target, the same target in each period needs to be tracked; based on the working characteristics of the security radar, the radar is static, is a static environment target in most cases, and has fewer moving targets, so that the targets are distinguished according to the moving and static states, the static target tracking process is simplified, and the moving targets can use a complex self-adaptive multi-model interactive tracking algorithm (IMM algorithm), thereby considering high precision and maneuverability;

to this end, in a first aspect, an embodiment of the present invention provides a feature-based target tracking optimization method, including the following steps:

step S10, processing the obtained target trace information to respectively form a static target trace and a moving target trace; performing static target data association to obtain a static target associated with a static target track; performing moving target data association to obtain a moving target associated with a moving target track;

and step S20, performing moving target tracking by utilizing an IMM algorithm based on the moving target related to the moving target track.

The specific processes of the above steps S10, S20 are explained in detail below;

referring to fig. 1, after obtaining the target trace information through the signal processing unit, step S10 specifically includes:

step S11, clustering the targets;

step S12, separating the moving and static targets according to the clustered speed information and characteristic information;

step S13, forming a static target track after the separated static target is initialized;

step S14, performing static target data association according to the static target association parameters and the association conditions to obtain a static target associated with the static target track;

step S15, the separated moving target forms a moving target track after being initialized;

step S16, performing moving target data association according to the moving target association parameters and the association conditions to obtain a moving target associated with a moving target track;

step S20 is performed after steps S14 and S16; namely, performing action target tracking by utilizing an IMM algorithm;

in the above process, because the static target itself is not likely to change in orientation, only the correlation operation is performed without tracking in order to save the algorithm time; the difficulty of the tracking algorithm is the uncertainty of the motion mode of the target; for a moving target, if the motion state of the adopted motion model is not matched with the motion state and the set model of the actual target, the estimation accuracy of the filter is reduced, the filter is diverged, and even the situations of track breakage and target loss occur; therefore, the embodiment of the invention adopts an IMM algorithm to track the moving target; model probabilities among the models are adaptively adjusted by utilizing an interactive multi-model algorithm, and a real motion track of the target is fitted; the adaptive multi-model interactive tracking algorithm (IMM algorithm) comprises the following target motion models: a uniform velocity model, a uniform acceleration model, a Singer model and a current statistical model; for a target in conventional motion, a uniform velocity model and a uniform acceleration model can be well fitted; for targets with high maneuverability performance such as turning, sudden stop and the like, the Singer model and the current statistical model are more suitable; the IMM algorithm treats the system as a Markov chain in a finite state, different models are interacted through transition probability, model transition conforms to the Markov process, a plurality of models can exist simultaneously, and switching is performed according to data to obtain better tracking performance;

referring to fig. 2, step S20 specifically includes:

step S21, inputting the posterior probability, Markov transition probability and covariance matrix of each model obtained by calculation in the previous period as input, and the information of the moving target related to the moving target track into each model;

here, the posterior probability of each model refers to the probability assigned to each model calculated from the measurement value in the previous cycle; the Markov transition probability is an empirical value matrix initially determined according to the conversion possibility among the models;

step S22, carrying out model information interaction: comprehensively calculating according to the posterior probability of each model and the Markov transition probability to obtain a mixed probability; initializing information of a moving target related to a moving target track and a covariance matrix through mixed probability to obtain interactive target information and an interactive covariance matrix;

step S23, interactive target information and an interactive covariance matrix of each model are used as filter input, and the filter of each model is input for parallel filtering;

step S24, obtaining outputs of the respective models in the current period: an estimated value and a covariance matrix;

step S25, model probability updating: calculating the filtering residual error of each model at the current moment by adopting a Bayesian hypothesis test method, and calculating the distribution probability of each model according to the estimated value and the measured value;

step S26, state result interaction: and performing weighted fusion on the estimation value and the covariance matrix of each model according to the distribution probability to obtain a final estimation result of the current period.

In a second aspect, an embodiment of the present invention further provides a feature-based target tracking optimization apparatus, including: a processor and a memory; the processor and the memory are communicated with each other; the memory has stored therein a computer program; the processor comprises a data processing unit for executing the computer program, which computer program when executed performs the steps of the method as described above. The Processor may be a CPU, or other general-purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or the like.

In a third aspect, an embodiment of the present invention also proposes a storage medium having stored therein a computer program configured to perform the steps of the method as described above when executed. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk (Hard Disk Drive, abbreviated as HDD) or a Solid State Drive (SSD), etc.; the storage medium may also comprise a combination of memories of the kind described above.

Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention has been described in detail with reference to examples, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.