Target Tracking Method, Device and Equipment Combining Uncertain Regions

Abstract

The present application provides a target tracking method, device and equipment combined with an uncertain area. The method includes: if the observation information of the target is obtained at the current update moment, the Kalman filter state estimation result is updated according to the obtained observation information; otherwise, through a preset state model, the most recent Kalman filter state estimation result is projected from the corresponding moment to the current update moment to obtain a new state vector and state covariance matrix. The center of the prediction area is determined according to the position coordinates therein, and the major axis length, minor axis length and inclination angle of the prediction area are determined according to the position sub-matrix therein. The preset state model assumes that the target moves in a straight line within a preset time period, and the movement speed is affected by exponential decay and random acceleration perturbation. Through the present application, after the observation information is lost, a prediction area with a smaller area can be given, and the probability that the target is within the prediction area is higher, significantly improving the confidence of the lost target estimation.

Description

Technical Field

[0001] This application relates to the technical field of target tracking, and particularly relates to a target tracking method, device, and equipment that combine an uncertain region. Background Art

[0002] In the field of target tracking, Kalman filtering and its improved algorithms (such as extended Kalman filtering, iterated extended Kalman filtering, etc.) are widely used in the prediction and estimation of the target's motion state. Existing methods describe the target's state and its uncertainty through a state vector and a state covariance matrix, and achieve precise tracking of the target through a "prediction-update" loop when the observation information is continuous. However, in a marine environment, due to factors such as sea wave noise, biological noise, and ship navigation noise, the observation information may be lost. When existing methods perform position prediction after the observation information is lost, they often can only give a prediction area with a large area, and the probability of the target being in this prediction area is low, which brings great difficulties and uncertainties to the actual target search and tracking work. Summary of the Invention

[0003] This application provides a target tracking method, device, and equipment that combine an uncertain region, which can solve the technical problem of low accuracy of the prediction area given after the observation information is lost in the existing technology.

[0004] In a first aspect, an embodiment of this application provides a target tracking method that combines an uncertain region. The target tracking method includes:

[0005] If the observation information of the target is obtained at the current update moment, the Kalman filter state estimation result is updated according to the obtained observation information;

[0006] If the observation information of the target is not obtained at the current update moment, the most recent Kalman filter state estimation result is used as the first state vector and the first state covariance matrix, and through a preset state model, the first state vector and the first state covariance matrix are projected from the corresponding moment to the current update moment to obtain a second state vector and a second state covariance matrix, where the preset state model assumes that the target moves in a straight line within a preset time period, and the motion speed is affected by exponential decay and random acceleration perturbation;

[0007] The center of the prediction area is determined according to the position coordinates in the second state vector, and the major axis length, minor axis length, and tilt angle of the prediction area are determined according to the position sub-matrix in the second state covariance matrix.

[0008] Further, in one embodiment, the state transition matrix of the preset state model includes a first parameter and a second parameter, and the noise matrix of the preset state model includes a third parameter, a fourth parameter, and a fifth parameter;

[0009] The first parameter is used to couple speed and position;

[0010] The second parameter is used to control the exponential decay of speed;

[0011] The third parameter is used to control the cumulative effect of position noise, including compensation for time integration and dynamic decay;

[0012] The fourth parameter is used to control the coupling effect of speed noise on position;

[0013] The fifth parameter is used to control the variance of speed noise;

[0014] The third parameter, the fourth parameter, and the fifth parameter are directly proportional to the square of the motion speed in the first state vector and inversely related to the preset duration.

[0015] Furthermore, in one embodiment, the state vector includes position information in the x direction, position information in the y direction, speed information in the x direction, and speed information in the y direction;

[0016] The step of projecting the first state vector and the first state covariance matrix from the corresponding moment to the current update moment through the preset state model to obtain the second state vector and the second state covariance matrix includes:

[0017] Calculate the second state vector and the second state covariance matrix according to the first formula. The first formula is:

[0018]

[0019]

[0020] Wherein, is the second state vector, is the second state covariance matrix, is the first state vector, is the first state covariance matrix, is the state transition matrix of the preset state model, is the noise matrix of the preset state model,

[0021] ,

[0022] , ,

[0023] ,

[0024] , , , ,

[0025] Among them, s is the moving speed in the first state vector, which is obtained by calculating the vector sum and taking the modulus of the x-direction speed information and the y-direction speed information in the first state vector. is the preset duration. is the time difference from the corresponding moment of the first state vector and the first state covariance matrix to the current update moment.

[0026] Further, in one embodiment, .

[0027] Further, in one embodiment, the state vector includes x-direction position information, y-direction position information, x-direction speed information, and y-direction speed information.

[0028] The steps of determining the major axis length, minor axis length, and tilt angle of the prediction region according to the position submatrix in the second state covariance matrix include:

[0029] Calculating the major axis length, minor axis length, and tilt angle of the prediction region according to the second formula, and the second formula is:

[0030] ,

[0031] ,

[0032] ,

[0033] Among them, Major, Minor, and EllipseBearing are the major axis length, minor axis length, and tilt angle of the prediction region respectively, a is the uncertainty degree of the x-direction position information in the second state covariance matrix, b is the uncertainty degree of the y-direction position information in the second state covariance matrix, and h is the correlation degree of the x-direction position information and the y-direction position information in the second state covariance matrix.

[0034] Further, in one embodiment, the observation information of the target is obtained through two observation stations.

[0035] The steps of updating the Kalman filter state estimation result according to the obtained observation information include:

[0036] Calculating the prior state vector and the prior state covariance matrix according to the third formula, and the third formula is:

[0037] ,

[0038] ,

[0039] Among them, is the prior state vector, is the prior state covariance matrix, is the state vector in the most recent Kalman filter state estimation result, is the state transition function, is the Jacobian matrix of, is the process noise covariance matrix;

[0040] The posterior state vector and the posterior state covariance matrix are calculated according to the fourth formula, and the fourth formula is:

[0041] ,

[0042] ,

[0043] ,

[0044] wherein, is the Kalman gain updated at the i-th iteration, is the posterior state vector updated at the i-th iteration, is the posterior state covariance matrix updated at the i-th iteration, , , is the observation information used in the i-th iteration update. If only the observation information of a single observation station is obtained, the observation information of this observation station is used for each iteration update. If the observation information of two observation stations is obtained, the observation information of the two observation stations is alternately used for iteration update, is the observation transfer function, is the Jacobian matrix of, is the observation noise covariance matrix;

[0045] If the number of iteration updates has not reached the upper limit, and the difference between the posterior state vectors of two adjacent iteration updates is greater than or equal to the difference threshold, then return to execute the step of calculating the posterior state vector and the posterior state covariance matrix according to the fourth formula;

[0046] If the number of iteration updates reaches the upper limit, or the difference between the posterior state vectors of two adjacent iteration updates is less than the difference threshold, then stop the iteration update, and use the posterior state vector and the posterior state covariance matrix of the last iteration update as the new Kalman filter state estimation result.

[0047] Furthermore, in one embodiment, after the step of stopping the iteration update, it further includes:

[0048] Update the observation noise covariance matrix according to the fifth formula, and the fifth formula is:

[0049] ,

[0050] ,

[0051] where m is the number of iterative updates, is the forgetting factor, .

[0052] Furthermore, in one embodiment, .

[0053] In a second aspect, an embodiment of the present application further provides a target tracking device combined with an uncertain region. The target tracking device includes:

[0054] A Kalman filter module, configured to update the Kalman filter state estimation result according to the obtained observation information if the observation information of the target is obtained at the current update moment;

[0055] A random motion module, configured to use the most recent Kalman filter state estimation result as the first state vector and the first state covariance matrix if the observation information of the target is not obtained at the current update moment, and project the first state vector and the first state covariance matrix from the corresponding moment to the current update moment through a preset state model to obtain a second state vector and a second state covariance matrix, where the preset state model assumes that the target moves linearly within a preset duration, and the motion speed is affected by exponential decay and random acceleration perturbation;

[0056] A region solving module, configured to determine the center of the prediction region according to the position coordinates in the second state vector, and determine the major axis length, minor axis length, and tilt angle of the prediction region according to the position sub-matrix in the second state covariance matrix.

[0057] In a third aspect, an embodiment of the present application further provides a target tracking device combined with an uncertain region. The target tracking device includes a processor, a memory, and a target tracking program stored on the memory and executable by the processor. When the target tracking program is executed by the processor, the steps of the above target tracking method are implemented.

[0058] In the present application, when the observation information of the target is not obtained, it is assumed that the target moves linearly within a preset duration, and the motion speed is affected by exponential decay and random acceleration perturbation. Based on the most recent Kalman filter state estimation result, the target state is further predicted, and an estimation result closer to the true state of the target can be obtained. The predicted region given is elliptical, the center position is more accurate, and the expansion speed is slower over time. Through the present application, after the observation information is lost, a prediction region with a smaller area can be given, and the probability that the target is within the prediction region is higher, significantly improving the confidence of the lost target estimation and facilitating the development of target search and tracking work. Description of the Drawings

[0059] Figure 1 It is a schematic flowchart of the target tracking method in an embodiment of the present application;

[0060] Figure 2 It is a schematic diagram comparing the Kalman filter state estimation trajectory, the true trajectory, and the observation trajectory in the simulation experiment;

[0061] Figure 3 It is a schematic diagram comparing the root mean square error between the Kalman filter state estimation result and the observation information in the simulation experiment;

[0062] Figure 4 It is a schematic diagram of the functional modules of the target tracking device in an embodiment of the present application;

[0063] Figure 5 It is a schematic diagram of the hardware structure of the target tracking device involved in the solution of the embodiment of the present application. Detailed implementation manners

[0064] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0065] To make the purpose, technical solution, and advantages of the present application clearer, the embodiments of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0066] In a first aspect, an embodiment of the present application provides a target tracking method combined with an uncertain region.

[0067] Figure 1 It shows a schematic flowchart of the target tracking method in an embodiment of the present application.

[0068] Referring to Figure 1 , in an embodiment, the target tracking method includes the following steps:

[0069] S1. If the observation information of the target is obtained at the current update moment, update the Kalman filter state estimation result according to the obtained observation information.

[0070] Optionally, the observation information of the target can be obtained through a single observation station or through two observation stations.

[0071] The accuracy of single - station pure - bearing tracking needs to be improved because it only relies on the bearing data of a single observation station. Due to the lack of multi - perspective information supplementation and verification, the obtained target position information often has a large deviation and is difficult to meet the requirements of high - precision tracking.

[0072] Compared with the single - station method, the dual - station can reduce the influence of measurement errors to a certain extent and improve the estimation accuracy of the target position. The bearing information of the two observation stations complements and corroborates each other, which helps to more accurately determine the position and motion trajectory of the target. Through the collaborative work of the two observation stations, the accuracy and stability of target positioning can be effectively improved, and the tracking error caused by the limitations of a single observation station can be reduced.

[0073] It should be noted that the Kalman filter in this application generally refers to the Kalman filter series algorithms, rather than a specific one. The appropriate algorithm can be selected from the Kalman filter series algorithms according to the application scenario to calculate the Kalman filter state estimation result. For example, the Extended Kalman Filter, Iterated Extended Kalman Filter, Adaptive Kalman Filter, etc.

[0074] S2. If the observation information of the target is not obtained at the current update moment, the most recent Kalman filter state estimation result is used as the first state vector and the first state covariance matrix. Through a preset state model, the first state vector and the first state covariance matrix are projected from the corresponding moment to the current update moment to obtain the second state vector and the second state covariance matrix. Among them, the preset state model assumes that the target moves in a straight line within a preset time period, and the moving speed is affected by exponential decay and random acceleration perturbation.

[0075] In this embodiment, when the observation information of the target is not obtained, assuming that the target moves in a straight line within a preset time period and the moving speed is affected by exponential decay and random acceleration perturbation, further predicting the target state based on the most recent Kalman filter state estimation result can obtain an estimation result closer to the true state of the target.

[0076] S3. Determine the center of the prediction area according to the position coordinates in the second state vector, and determine the major - axis length, minor - axis length, and tilt angle of the prediction area according to the position sub - matrix in the second state covariance matrix.

[0077] In the existing scheme, when the observation information of the target is not obtained, a circular prediction area is given. The center is the position in the most recent Kalman filter state estimation result, and the radius is the speed in the most recent Kalman filter state estimation result multiplied by the time difference from the corresponding moment to the current update moment. This prediction area will expand rapidly over time, the center position will not change with time, and the accuracy will become lower and lower.

[0078] In this embodiment, the prediction region is elliptical, with a smaller range compared to a circular shape, a slower expansion speed over time, and the center position changing over time, resulting in higher accuracy.

[0079] Therefore, in this embodiment, when no observation information of the target is obtained, it is assumed that the target moves linearly within a preset time duration, and the movement speed is affected by exponential decay and random acceleration perturbations. Based on the most recent Kalman filter state estimation result, the target state is further predicted, and an estimation result closer to the true state of the target can be obtained. The given prediction region is elliptical, the center position is more accurate, and the expansion speed over time is slower. Through this embodiment, after the observation information is lost, a prediction region with a smaller area can be given, and the probability that the target is within this prediction region is higher, significantly improving the confidence level of the lost target estimation and facilitating the carrying out of target search and tracking work.

[0080] Furthermore, in one embodiment, the state transition matrix of the preset state model includes a first parameter and a second parameter, and the noise matrix of the preset state model includes a third parameter, a fourth parameter, and a fifth parameter;

[0081] The first parameter is used to couple the speed and the position;

[0082] The second parameter is used to control the exponential decay of the speed;

[0083] The third parameter is used to control the cumulative effect of the position noise, including the compensation of time integration and dynamic decay;

[0084] The fourth parameter is used to control the coupling effect of the speed noise on the position;

[0085] The fifth parameter is used to control the variance of the speed noise;

[0086] The third parameter, the fourth parameter, and the fifth parameter are proportional to the square of the movement speed in the first state vector and are negatively correlated with the preset time duration.

[0087] Furthermore, in one embodiment, the state vector includes the x-direction position information, the y-direction position information, the x-direction speed information, and the y-direction speed information;

[0088] The step of projecting the first state vector and the first state covariance matrix from the corresponding moment to the current update moment through the preset state model to obtain the second state vector and the second state covariance matrix includes:

[0089] Calculating the second state vector and the second state covariance matrix according to the first formula, and the first formula is:

[0090]

[0091]

[0092] Among them, is the second state vector, is the second state covariance matrix, is the first state vector, is the first state covariance matrix, is the state transition matrix of the preset state model, is the noise matrix of the preset state model,

[0093] ,

[0094] , ,

[0095] ,

[0096] , , , ,

[0097] Among them, s is the motion speed in the first state vector, which is obtained by calculating the vector sum and taking the modulus of the x-direction speed information and the y-direction speed information in the first state vector. is the preset duration, is the time difference from the corresponding moment of the first state vector and the first state covariance matrix to the current update moment.

[0098] It can be seen that b1, b2, c1, c2, and c3 are respectively the first parameter, the second parameter, the third parameter, the fourth parameter, and the fifth parameter described above.

[0099] For simplicity of calculation, it can be set that .

[0100] Furthermore, in one embodiment, the state vector includes x-direction position information, y-direction position information, x-direction speed information, and y-direction speed information;

[0101] The steps of determining the major axis length, minor axis length, and tilt angle of the prediction region according to the position sub-matrix in the second state covariance matrix include:

[0102] Calculating the major axis length, minor axis length, and tilt angle of the prediction region according to the second formula. The second formula is:

[0103] ,

[0104] ,

[0105] ,

[0106] Among them, Major, Minor, and EllipseBearing are respectively the major axis length, minor axis length, and tilt angle of the prediction region, a is the uncertainty degree of the x-direction position information in the second state covariance matrix, b is the uncertainty degree of the y-direction position information in the second state covariance matrix, and h is the correlation degree between the x-direction position information and the y-direction position information in the second state covariance matrix.

[0107] Further, in one embodiment, the observation information of the target is obtained through two observation stations;

[0108] The step of updating the Kalman filter state estimation result according to the obtained observation information includes:

[0109] The prior state vector and the prior state covariance matrix are calculated according to the third formula, and the third formula is:

[0110] ,

[0111] ,

[0112] Among them, is the prior state vector, is the prior state covariance matrix, is the state vector in the most recent Kalman filter state estimation result, is the state transition function, is the Jacobian matrix of is the process noise covariance matrix;

[0113] The posterior state vector and the posterior state covariance matrix are calculated according to the fourth formula, and the fourth formula is:

[0114] ,

[0115] ,

[0116] ,

[0117] Among them, is the Kalman gain updated in the i-th iteration, is the posterior state vector updated in the i-th iteration, is the posterior state covariance matrix updated in the i-th iteration, , , The observation information used for the i-th iterative update. If only the observation information of a single observation station is obtained, the observation information of this observation station is used for each iterative update. If the observation information of two observation stations is obtained, the observation information of the two observation stations is alternately used for iterative update. is the observation transition function. is for the Jacobian matrix of is the observation noise covariance matrix;

[0118] If the number of iterative updates has not reached the upper limit, and the difference between the posterior state vectors of two adjacent iterative updates is greater than or equal to the difference threshold, then return to execute the step of calculating the posterior state vector and the posterior state covariance matrix according to the fourth formula;

[0119] If the number of iterative updates reaches the upper limit, or the difference between the posterior state vectors of two adjacent iterative updates is less than the difference threshold, then stop the iterative update, and use the posterior state vector and the posterior state covariance matrix of the last iterative update as the new Kalman filter state estimation result.

[0120] In this embodiment, the posterior estimation link in the Kalman filter adopts an iterative method, making full use of the observation information obtained from two observation stations, so as to obtain an accurate Kalman filter state estimation result.

[0121] Further, in one embodiment, after the step of stopping the iterative update, it further includes:

[0122] Update the observation noise covariance matrix according to the fifth formula, and the fifth formula is:

[0123] ,

[0124] ,

[0125] where m is the number of iterative updates, is the forgetting factor, .

[0126] In this embodiment, after stopping the iterative update, a new observation noise covariance matrix is obtained according to the current observation noise covariance matrix and the residual situation of the last two iterative updates, so as to effectively cope with the change of the observation noise and enhance the adaptability to the complex environment.

[0127] According to experience, it can be set that .

[0128] Figure 2 shows a comparison diagram of the Kalman filter state estimation trajectory, the true trajectory and the observation trajectory in the simulation experiment. Figure 3It shows a schematic diagram of the comparison of the root mean square error between the Kalman filter state estimation result and the observation information in the simulation experiment.

[0129] It is verified through a two-dimensional uniform linear motion simulation experiment.

[0130] Referring to Figure 2 , the red line is the real target motion trajectory, the blue line is the motion trajectory with observation noise, and the green line is the Kalman filter state estimation trajectory in this embodiment. It can be seen that the Kalman filter operation in this embodiment effectively reduces the error in the observation and can provide a position closer to the real value.

[0131] Referring to Figure 3 , the red line and the green line are respectively the root mean square errors of the observation information of two observation stations, and the blue line is the root mean square error of the Kalman filter state estimation result in this embodiment. It can be seen that the Kalman filter operation in this embodiment can effectively reduce the root mean square error of the target in the azimuth.

[0132] In a second aspect, the embodiment of the present application further provides a target tracking device combined with an uncertain region.

[0133] Figure 4 It shows a schematic diagram of the functional modules of the target tracking device in an embodiment of the present application.

[0134] Referring to Figure 4 , in an embodiment, the target tracking device includes:

[0135] A Kalman filter module 10, configured to update the Kalman filter state estimation result according to the obtained observation information if the observation information of the target is obtained at the current update moment;

[0136] A random motion module 20, configured to use the most recent Kalman filter state estimation result as the first state vector and the first state covariance matrix if the observation information of the target is not obtained at the current update moment, and project the first state vector and the first state covariance matrix from the corresponding moment to the current update moment through a preset state model to obtain a second state vector and a second state covariance matrix, where the preset state model assumes that the target moves linearly within a preset time period, and the motion speed is affected by exponential decay and random acceleration perturbation;

[0137] A region solving module 30, configured to determine the center of the prediction region according to the position coordinates in the second state vector, and determine the major axis length, minor axis length, and tilt angle of the prediction region according to the position sub-matrix in the second state covariance matrix.

[0138] Further, in an embodiment, the state transition matrix of the preset state model includes a first parameter and a second parameter, and the noise matrix of the preset state model includes a third parameter, a fourth parameter, and a fifth parameter;

[0139] The first parameter is used to couple speed with position;

[0140] The second parameter is used to control the exponential decay of speed;

[0141] The third parameter is used to control the cumulative effect of position noise, including compensation for time integration and dynamic decay;

[0142] The fourth parameter is used to control the coupling effect of speed noise on position;

[0143] The fifth parameter is used to control the variance of speed noise;

[0144] The third parameter, the fourth parameter, and the fifth parameter are directly proportional to the square of the moving speed in the first state vector and inversely related to the preset duration.

[0145] Furthermore, in one embodiment, the state vector includes position information in the x direction, position information in the y direction, speed information in the x direction, and speed information in the y direction;

[0146] The random motion module 20 is used for:

[0147] Calculating a second state vector and a second state covariance matrix according to a first formula, and the first formula is:

[0148]

[0149]

[0150] Wherein, is the second state vector, is the second state covariance matrix, is the first state vector, is the first state covariance matrix, is the state transition matrix of the preset state model, is the noise matrix of the preset state model,

[0151] ,

[0152] , ,

[0153] ,

[0154] , , , ,

[0155] Among them, s is the motion speed in the first state vector, which is obtained by calculating the vector sum and taking the modulus of the x-direction speed information and the y-direction speed information in the first state vector. is a preset time duration. is the time difference from the corresponding moment of the first state vector and the first state covariance matrix to the current update moment.

[0156] Further, in one embodiment, .

[0157] Further, in one embodiment, the state vector includes x-direction position information, y-direction position information, x-direction speed information, and y-direction speed information.

[0158] The region solving module 30 is used for:

[0159] Calculating the major axis length, minor axis length, and tilt angle of the predicted region according to the second formula, and the second formula is:

[0160] ,

[0161] ,

[0162] ,

[0163] Among them, Major, Minor, and EllipseBearing are the major axis length, minor axis length, and tilt angle of the predicted region respectively, a is the uncertainty degree of the x-direction position information in the second state covariance matrix, b is the uncertainty degree of the y-direction position information in the second state covariance matrix, and h is the correlation degree between the x-direction position information and the y-direction position information in the second state covariance matrix.

[0164] Further, in one embodiment, the observation information of the target is obtained through two observation stations.

[0165] The Kalman filtering module 10 is used for:

[0166] Calculating the prior state vector and the prior state covariance matrix according to the third formula, and the third formula is:

[0167] ,

[0168] ,

[0169] Among them, is the prior state vector, is the prior state covariance matrix, is the state vector in the nearest Kalman filtering state estimation result, is the state transition function, For the Jacobian matrix, is the process noise covariance matrix;

[0170] The posterior state vector and the posterior state covariance matrix are calculated according to the fourth formula, and the fourth formula is:

[0171] ,

[0172] ,

[0173] ,

[0174] wherein, is the Kalman gain updated in the i-th iteration, is the posterior state vector updated in the i-th iteration, is the posterior state covariance matrix updated in the i-th iteration, , , is the observation information used in the i-th iteration update. If only the observation information of a single observation station is obtained, the observation information of this observation station is used for each iteration update. If the observation information of two observation stations is obtained, the observation information of the two observation stations is alternately used for iteration update, is the observation transfer function, is the Jacobian matrix of is the observation noise covariance matrix;

[0175] If the number of iteration updates does not reach the upper limit, and the difference between the posterior state vectors of two adjacent iteration updates is greater than or equal to the difference threshold, then return to execute the step of calculating the posterior state vector and the posterior state covariance matrix according to the fourth formula;

[0176] If the number of iteration updates reaches the upper limit, or the difference between the posterior state vectors of two adjacent iteration updates is less than the difference threshold, then stop the iteration update, and use the posterior state vector and the posterior state covariance matrix of the last iteration update as the new Kalman filter state estimation result.

[0177] Furthermore, in an embodiment, the Kalman filter module 10 is further configured to:

[0178] After the step of stopping the iteration update, update the observation noise covariance matrix according to the fifth formula, and the fifth formula is:

[0179] ,

[0180] ,

[0181] Among them, m is the number of iterative updates, is the forgetting factor, .

[0182] Furthermore, in one embodiment, .

[0183] Among them, the function implementation of each module in the above target tracking device corresponds to each step in the above target tracking method embodiment, and its function and implementation process will not be elaborated here one by one.

[0184] In a third aspect, an embodiment of the present application provides a target tracking device combined with an uncertain region. The target tracking device can be a device with data processing functions such as a personal computer (PC), a laptop computer, a server, etc.

[0185] Figure 5 shows a schematic hardware structure diagram of the target tracking device involved in the embodiment solution of the present application.

[0186] Referring to Figure 5 , in the embodiment of the present application, the target tracking device may include a processor, a memory, a communication interface, and a communication bus.

[0187] Among them, the communication bus can be of any type and is used to interconnect the processor, the memory, and the communication interface.

[0188] The communication interface includes interfaces such as input / output (I / O) interfaces, physical interfaces, and logical interfaces for interconnecting internal devices of the target tracking device, as well as interfaces for interconnecting the target tracking device with other devices (such as other computing devices or user devices). The physical interface can be an Ethernet interface, a fiber optic interface, an ATM interface, etc.; the user device can be a display screen (Display), a keyboard (Keyboard), etc.

[0189] The memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical memory, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0190] The processor may be a general-purpose processor, which can call the target tracking program stored in the memory and execute the target tracking method provided in the embodiments of the present application. For example, the general-purpose processor may be a central processing unit (CPU). Wherein, the method executed when the target tracking program is called may refer to the various embodiments of the target tracking method of the present application, which will not be elaborated here.

[0191] Those skilled in the art can understand that Figure 5 the hardware structure shown in does not constitute a limitation to the present application, and may include more or fewer components than shown in the figure, or combine certain components, or arrange different components.

[0192] It should be noted that the serial numbers of the embodiments of the present application above are only for description and do not represent the superiority or inferiority of the embodiments.

[0193] The terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices. The descriptions with terms such as "first", "second" and "third" are used to distinguish different objects, etc., which do not represent a sequence, nor do they limit that "first", "second" and "third" are different types.

[0194] In the description of the embodiments of the present application, "exemplary", "for example" or "for instance" etc. are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "exemplary", "for example" or "for instance" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of words such as "exemplary", "for example" or "for instance" is intended to present relevant concepts in a specific manner.

[0195] In the description of the embodiments of the present application, unless otherwise specified, " / " means "or". For example, A / B may mean A or B; "and / or" in the text is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B may mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of the present application, "a plurality of" means two or more than two.

[0196] In some of the processes described in the embodiments of this application, there are multiple operations or steps that appear in a specific order. However, it should be understood that these operations or steps may not be executed in the order in which they appear in the embodiments of this application or may be executed in parallel. The serial numbers of the operations are only used to distinguish different operations, and the serial numbers themselves do not represent any order of execution. Additionally, these processes may include more or fewer operations, and these operations or steps may be executed in sequence or in parallel, and these operations or steps may be combined.

[0197] Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases, the former is a better implementation method. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium as described above (such as ROM / RAM, magnetic disk, optical disc) and includes several instructions to enable a terminal device to execute the methods described in various embodiments of this application.

[0198] The above are only the preferred embodiments of this application, and do not limit the patent scope of this application. Any equivalent structural or equivalent process transformation made by using the content of the specification and drawings of this application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of this application.

Claims

1. A target tracking method combined with an uncertain region, characterized in that, The target tracking method includes: If the observation information of the target is obtained at the current update moment, the Kalman filter state estimation result is updated according to the obtained observation information; If the observation information of the target is not obtained at the current update moment, the most recent Kalman filter state estimation result is used as the first state vector and the first state covariance matrix, and through a preset state model, the first state vector and the first state covariance matrix are projected from the corresponding moment to the current update moment to obtain a second state vector and a second state covariance matrix, where the preset state model assumes that the target moves linearly within a preset duration, and the moving speed is affected by exponential decay and random acceleration perturbation; The center of the prediction region is determined according to the position coordinates in the second state vector, and the major axis length, minor axis length, and tilt angle of the prediction region are determined according to the position sub-matrix in the second state covariance matrix.

2. The target tracking method according to claim 1, wherein The state transition matrix of the preset state model includes a first parameter and a second parameter, and the noise matrix of the preset state model includes a third parameter, a fourth parameter, and a fifth parameter; The first parameter is used to couple the speed and the position; The second parameter is used to control the exponential decay of the speed; The third parameter is used to control the cumulative effect of the position noise, including time integration and compensation for dynamic decay; The fourth parameter is used to control the coupling effect of the speed noise on the position; The fifth parameter is used to control the variance of the speed noise; The third parameter, the fourth parameter, and the fifth parameter are proportional to the square of the moving speed in the first state vector and are negatively correlated with the preset duration.

3. The target tracking method according to claim 1, wherein, The state vector includes x-direction position information, y-direction position information, x-direction speed information, and y-direction speed information; The step of projecting the first state vector and the first state covariance matrix from the corresponding moment to the current update moment through the preset state model to obtain a second state vector and a second state covariance matrix includes: The second state vector and the second state covariance matrix are calculated according to the first formula, and the first formula is: Among them, is the second state vector, is the second state covariance matrix, is the first state vector, is the first state covariance matrix, is the state transition matrix of the preset state model, is the noise matrix of the preset state model, ， ， ， ， ， ， ， ， Wherein, s is the movement speed in the first state vector, which is obtained by calculating the vector sum and taking the modulus of the x-direction speed information and the y-direction speed information in the first state vector. is the preset duration. is the time difference from the corresponding moment of the first state vector and the first state covariance matrix to the current update moment.

4. The target tracking method according to claim 3, wherein 。 5. The target tracking method according to claim 1, characterized in that, The state vector includes x-direction position information, y-direction position information, x-direction speed information, and y-direction speed information; The step of determining the major axis length, minor axis length, and tilt angle of the prediction region according to the position sub-matrix in the second state covariance matrix includes: The major axis length, minor axis length, and tilt angle of the prediction region are calculated according to the second formula, and the second formula is: ， ， ， Among them, Major, Minor, and EllipseBearing are the major axis length, minor axis length, and tilt angle of the prediction region respectively, a is the uncertainty degree of the x-direction position information in the second state covariance matrix, b is the uncertainty degree of the y-direction position information in the second state covariance matrix, and h is the correlation degree between the x-direction position information and the y-direction position information in the second state covariance matrix.

6. The target tracking method according to claim 1, characterized in that The observation information of the target is obtained through two observation stations; The step of updating the Kalman filter state estimation result according to the obtained observation information includes: The prior state vector and the prior state covariance matrix are calculated according to the third formula, and the third formula is: ， ， Among them, is the prior state vector, is the prior state covariance matrix, is the state vector in the most recent Kalman filter state estimation result, is the state transition function, is the Jacobian matrix of is the process noise covariance matrix; The posterior state vector and the posterior state covariance matrix are calculated according to the fourth formula, and the fourth formula is: ， ， ， Among them, is the Kalman gain updated at the i-th iteration, is the posterior state vector updated at the i-th iteration, is the posterior state covariance matrix updated at the i-th iteration, , , is the observation information used for the i-th iteration update. If only the observation information of a single observation station is obtained, the observation information of this observation station is used for each iteration update. If the observation information of two observation stations is obtained, the observation information of the two observation stations is alternately used for iterative update, is the observation transfer function, is the Jacobian matrix of is the observation noise covariance matrix; If the number of iterative updates has not reached the upper limit, and the difference between the posterior state vectors of two adjacent iterative updates is greater than or equal to the difference threshold, then return to execute the step of calculating the posterior state vector and the posterior state covariance matrix according to the fourth formula; If the number of iterative updates reaches the upper limit, or the difference between the posterior state vectors of two adjacent iterative updates is less than the difference threshold, then stop the iterative update, and use the posterior state vector and the posterior state covariance matrix of the last iterative update as the new Kalman filter state estimation result.

7. The target tracking method according to claim 6, characterized in that After the step of stopping the iterative update, it further includes: Updating the observation noise covariance matrix according to the fifth formula, and the fifth formula is: ， ， where m is the number of iterative updates, is the forgetting factor, .

8. The target tracking method according to claim 7, wherein 。 9. An object tracking device combined with an uncertain area, characterized in that, The target tracking device includes: A Kalman filter module, configured to update the Kalman filter state estimation result according to the obtained observation information if the observation information of the target is obtained at the current update moment; A random motion module, configured to, if the observation information of the target is not obtained at the current update moment, use the most recent Kalman filter state estimation result as the first state vector and the first state covariance matrix, and project the first state vector and the first state covariance matrix from the corresponding moment to the current update moment through a preset state model to obtain a second state vector and a second state covariance matrix, where the preset state model assumes that the target moves linearly within a preset time duration, and the motion speed is affected by exponential decay and random acceleration perturbations; A region solving module, configured to determine the center of the prediction region according to the position coordinates in the second state vector, and determine the major axis length, minor axis length, and tilt angle of the prediction region according to the position sub-matrix in the second state covariance matrix.

10. A target tracking device incorporating an uncertain region, characterized in that, The target tracking device includes a processor, a memory, and a target tracking program stored on the memory and executable by the processor. When the target tracking program is executed by the processor, the steps of the target tracking method according to any one of claims 1 to 8 are implemented.

Images