A method, device, electronic device and storage medium for determining heading angle

Abstract

The present application discloses a heading angle determination method, device, electronic device and storage medium, which are used to improve the accuracy of determining the heading angle of a moving vehicle. In the embodiment of the present application, a millimeter wave radar and a laser radar are first used to continuously detect the target scene respectively, and a first detection point cluster and a second detection point cloud are obtained; then the first detection point cluster and the second detection point cloud are subjected to time synchronization processing and space synchronization processing; and the coordinates of each first detection point after synchronization processing are regionally expanded to obtain an extended area of ​​the first detection point, and a second detection point falling in the extended area is determined; then the second detection point is associated with the first detection point, and the first detection point and the second detection point are information-fused to obtain a fused detection point; finally, the fused detection point cloud composed of the fused detection points is segmented to obtain a fused detection point cloud corresponding to each vehicle in the target scene, and the heading angle of each vehicle is determined based on the fused detection point cloud.

Description

Background Art

[0002] When estimating the motion state of a moving vehicle at the next moment, the heading angle of the moving vehicle is a very important parameter.

[0003] In the related technology, millimeter-wave radar or laser radar is mostly used to estimate the heading angle of a moving vehicle. When using millimeter-wave radar to estimate the heading angle of a moving vehicle, the millimeter-wave radar is first used to measure the position of the vehicle and its Doppler information, and then the corresponding heading angle estimation model is established in combination with the selected tracking model. However, the measurement accuracy of this method is limited by the error of the millimeter-wave radar in measuring the position of the moving vehicle, the error of the Doppler measurement, and the error of the heading angle estimation model; when using laser radar to estimate the heading angle of a moving vehicle, the point cloud obtained by detection is first segmented, and then the shape of the point cloud is estimated based on the point cloud to obtain the heading angle of the moving vehicle. However, the accuracy of the heading angle measurement of this method is limited by the accuracy of the point cloud obtained by detection. Summary of the invention

[0004] The purpose of the present application is to provide a heading angle determination method, device, electronic device and storage medium, which are used to improve the accuracy of determining the heading angle of a moving vehicle.

[0005] In a first aspect, an embodiment of the present application provides a method for determining a heading angle, the method comprising:

[0006] The target scene is continuously detected by a millimeter-wave radar to obtain a first detection point cluster of multiple consecutive frames, and the target scene is continuously detected by a laser radar to obtain a second detection point cloud of multiple consecutive frames; wherein the target scene includes at least one vehicle;

[0007] Performing time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain a first detection point cluster and a second detection point cloud that are associated in both time and space;

[0008] Expanding the coordinates of each first detection point in the first detection point cluster after synchronization according to the standard deviation to obtain an extended area of ​​the first detection point, and determining a second detection point falling in the extended area according to the coordinates of the second detection point; wherein the standard deviation is a standard deviation of a physical parameter associated with the first detection point;

[0009] Associating a second detection point falling in the extended area with the first detection point, and fusing information of the mutually associated first detection point and second detection point to obtain a fused detection point; the fused detection point constitutes a fused detection point cloud;

[0010] The fused detection point cloud formed by the fused detection points is segmented to obtain a fused detection point cloud corresponding to each vehicle in the target scene, and the heading angle of each vehicle is determined based on the fused detection point cloud.

[0011] In the present application, millimeter-wave radar and laser radar are used together to detect moving vehicles, and the first detection point cluster and the second detection point cloud obtained are associated, and the information of the first detection point and the corresponding second detection point are fused, so that the fused detection point has the information obtained by the millimeter-wave radar detection and the laser radar detection. Finally, the heading angle is estimated based on the fused detection point. Since the fused detection point has the physical parameter information of the millimeter-wave radar detection point and the density of the laser radar detection point, the heading angle can be more accurate.

[0012] In some possible embodiments, the performing time synchronization processing on the first detection point cluster and the second detection point cloud includes:

[0013] For each second timestamp, determine a first calculation timestamp and a second calculation timestamp corresponding to the second timestamp; wherein the first timestamp is obtained by marking a first detection point cluster of each frame obtained by continuously detecting the target scene by the millimeter wave radar at a first preset frequency, and the second timestamp is obtained by marking a second detection point cloud of each frame obtained by continuously detecting the target scene by the laser radar at a second preset frequency; wherein the first calculation timestamp is a first timestamp that is located before the second timestamp in time sequence and has the smallest time interval with the second timestamp, and the second calculation timestamp is a first timestamp that is located after the second timestamp in time sequence and has the smallest time interval with the second timestamp;

[0014] An associated first detection point cluster having an association relationship with a second detection point cloud corresponding to the second timestamp is determined based on the first calculation timestamp, the second calculation timestamp, and the second timestamp.

[0015] In the present application, the first detection point cluster and the second detection point cloud are time-synchronized based on the marked timestamps, and the first detection point cluster and the second detection point cloud that are temporally associated are determined, further improving the accuracy of subsequent determination of the heading angle.

[0016] In some possible embodiments, determining, based on the first calculation timestamp, the second calculation timestamp, and the second timestamp, an associated first detection point cluster having an association relationship with the second detection point cloud corresponding to the second timestamp includes:

[0017] Obtain a third timestamp, wherein the third timestamp is determined to be obtained by marking at a second preset frequency starting from a first time interval after the laser radar marks the first second timestamp; wherein the first time interval is obtained according to the second preset frequency;

[0018] For each third timestamp, determine a target second timestamp that has the smallest timing difference with the third timestamp and is before the third timestamp;

[0019] Obtain a first calculation timestamp and a second calculation timestamp corresponding to the target second timestamp;

[0020] Determine an associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp;

[0021] The associated first detection point cluster is used as a first detection point cluster that has a temporal association relationship with the second detection point cloud corresponding to the target second timestamp.

[0022] In the present application, the third timestamp is used to ensure that when calculating the first detection point cluster and the second detection point cloud having an associated relationship, the data of the first detection point cluster and the second detection point cloud have been detected, thereby ensuring the timeliness of the calculation.

[0023] In some possible embodiments, determining the associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp includes:

[0024] Performing linear interpolation processing on the detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster; or

[0025] Mean processing is performed on the coordinates of the first detection point in the detection point cluster corresponding to the first calculation timestamp and the first detection point in the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster.

[0026] In the present application, in order to ensure the accuracy of the final heading angle, the first detection point cluster corresponding to the first calculation timestamp and the second detection point cluster corresponding to the second calculation timestamp are combined to obtain an associated detection point cluster, so that the heading angle is more accurate.

[0027] In some possible embodiments, performing spatial synchronization processing on the first detection point cluster and the second detection point cloud includes:

[0028] The coordinates of the first detection point are transformed based on the rotation parameters and the translation parameters, and the coordinates of the second detection point are transformed based on the rotation parameters and the translation parameters to obtain the coordinates of the first detection point and the coordinates of the second detection point after spatial synchronization; wherein the rotation parameters are used to rotate the coordinate axes in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point to the left to coincide with the coordinate axes of the spatial synchronization coordinate system; and the translation parameters are used to determine the coordinate components of the origin of the spatial synchronization coordinate system in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point.

[0029] In the present application, the first detection point cloud and the second detection point cluster are spatially synchronized so that the coordinates of the first detection point in the first detection point cluster and the coordinates of the second detection point in the second detection point cloud can be expressed based on the same coordinate system.

[0030] In some possible embodiments, the rotation parameter and the translation parameter are obtained according to the following method:

[0031] Using a laser radar to measure a target object, obtain at least one measurement point of the target object and coordinates corresponding to the measurement point; and determine the mean value of the coordinates of the measurement point;

[0032] Measuring the target object using a millimeter wave radar to obtain a first coordinate of the target object;

[0033] A rotation parameter and a translation parameter are determined based on the mean of the coordinates and the first coordinate.

[0034] In the present application, the translation parameters and rotation parameters are determined based on the measurement values ​​of the target object by the millimeter wave radar and the laser radar, so that the spatial synchronization of the first detection point cluster and the second detection point cloud is more accurate.

[0035] In some possible embodiments, the coordinates of the first detection point include radial distance, azimuth angle and elevation angle;

[0036] The step of performing area expansion on the coordinates of each first detection point in the first detection point cluster after synchronization processing according to the standard deviation includes:

[0037] The following process is performed for each first detection point in the first detection point cluster:

[0038] Determine a radial distance standard deviation, an azimuth standard deviation, and an elevation standard deviation of the first detection point according to a signal-to-noise ratio of the first detection point;

[0039] Determine a first difference between the radial distance of the first detection point and the standard deviation of the radial distance, and determine a first sum between the radial distance of the first detection point and the standard deviation of the radial distance;

[0040] Determine an expansion area of ​​the radial distance according to the first difference and the first sum;

[0041] Determine a second difference between the azimuth of the first detection point and the azimuth standard deviation, and determine a second sum between the azimuth of the first detection point and the azimuth standard deviation;

[0042] Determine an extended area of ​​the azimuth according to the second difference and the second sum;

[0043] Determine a third difference between the pitch angle of the first detection point and the standard deviation of the pitch angle, and determine a third sum between the pitch angle of the first detection point and the standard deviation of the pitch angle;

[0044] Determining an extended area of ​​the pitch angle according to the third sum;

[0045] An expansion area of ​​the first detection point is determined according to the expansion area of ​​the radial distance, the expansion area of ​​the azimuth angle, and the expansion area of ​​the elevation angle.

[0046] In the embodiment of the present application, a method of expanding the area of ​​the first detection point is adopted to ensure that each second detection point has a corresponding first detection point.

[0047] In some possible embodiments, associating a second detection point that falls within a region corresponding to the first detection point with the first detection point includes:

[0048] If a plurality of second detection points fall within the extended area of ​​the first detection point, determining a second detection point with the largest signal-to-noise ratio in the extended area;

[0049] The second detection point with the largest signal-to-noise ratio is used as the second detection point associated with the first detection point.

[0050] In some possible embodiments, the first detection point cluster includes: a plurality of first detection points, the second detection point cloud includes: a plurality of second detection points, and the step of fusing information of the mutually related first detection points and second detection points to obtain fused detection points includes:

[0051] For each second detection point, perform the following process:

[0052] Using the physical parameter information of the first detection point associated with the second detection point as the physical parameter information of the fused detection point corresponding to the second detection point;

[0053] Using the spatially synchronized coordinates of the second detection point as the coordinates of the fused detection point corresponding to the second detection point;

[0054] The fused detection point is formed based on the physical parameter information of the fused detection point and the coordinates of the fused detection point.

[0055] In the present application, the physical parameter information of the first detection point and the coordinates of the second detection point are used so that the fused detection point can have the advantages of both the laser radar and the millimeter wave radar.

[0056] In some possible embodiments, determining the heading angle of each vehicle based on the fused detection point cloud includes:

[0057] For each frame of each vehicle, perform the fused detection point cloud:

[0058] Determining an initial heading angle of the vehicle based on the fused detection point cloud;

[0059] Filter out a target fusion detection point according to the signal-to-noise ratio of each fusion detection point in the fusion detection point cloud;

[0060] Determining a wheel position difference of the vehicle based on the fused detection point cloud and a fused detection point cloud of a previous frame;

[0061] The initial heading angle, the target fusion detection point and the wheel position difference are input into a pre-trained motion model to obtain the heading angle of the vehicle.

[0062] In the present application, the heading angle of the vehicle is obtained based on the fusion detection point cloud, thereby improving the accuracy of estimating the heading angle of the vehicle.

[0063] In a second aspect, the present application further provides a heading angle determination device, the device comprising:

[0064] A detection module, configured to continuously detect the target scene using a millimeter wave radar to obtain a first detection point cluster of multiple consecutive frames, and to continuously detect the target scene using a laser radar to obtain a second detection point cloud of multiple consecutive frames; wherein the target scene includes at least one vehicle;

[0065] A synchronization module, used for performing time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain a first detection point cluster and a second detection point cloud that are associated in both time and space;

[0066] An expansion module, configured to perform area expansion on the coordinates of each first detection point in the first detection point cluster after synchronization processing according to a standard deviation to obtain an extended area of ​​the first detection point, and determine a second detection point falling in the extended area according to the coordinates of the second detection point; wherein the standard deviation is a standard deviation of a physical parameter associated with the first detection point;

[0067] a fusion module, configured to associate a second detection point falling in the extended area with the first detection point, and to fuse information of the mutually associated first detection point and second detection point to obtain a fused detection point; the fused detection point constitutes a fused detection point cloud;

[0068] The heading angle determination module is used to segment the fused detection point cloud composed of the fused detection points to obtain the fused detection point cloud corresponding to each vehicle in the target scene, and determine the heading angle of each vehicle based on the fused detection point cloud.

[0069] In some possible embodiments, when the synchronization module performs time synchronization processing on the first detection point cluster and the second detection point cloud, it is configured as follows:

[0070] For each second timestamp, determine a first calculation timestamp and a second calculation timestamp corresponding to the second timestamp; wherein the first timestamp is obtained by marking a first detection point cluster of each frame obtained by continuously detecting the target scene by the millimeter wave radar at a first preset frequency, and the second timestamp is obtained by marking a second detection point cloud of each frame obtained by continuously detecting the target scene by the laser radar at a second preset frequency; wherein the first calculation timestamp is a first timestamp that is located before the second timestamp in time sequence and has the smallest time interval with the second timestamp, and the second calculation timestamp is a first timestamp that is located after the second timestamp in time sequence and has the smallest time interval with the second timestamp;

[0071] An associated first detection point cluster having an association relationship with a second detection point cloud corresponding to the second timestamp is determined based on the first calculation timestamp, the second calculation timestamp, and the second timestamp.

[0072] In some possible embodiments, when the synchronization module determines, based on the first calculation timestamp, the second calculation timestamp, and the second timestamp, an associated first detection point cluster having an association relationship with the second detection point cloud corresponding to the second timestamp, the synchronization module is configured as follows:

[0073] Obtain a third timestamp, wherein the third timestamp is determined to be obtained by marking at a second preset frequency starting from a first time interval after the laser radar marks the first second timestamp; wherein the first time interval is obtained according to the second preset frequency;

[0074] For each third timestamp, determine a target second timestamp that has the smallest timing difference with the third timestamp and is before the third timestamp;

[0075] Obtain a first calculation timestamp and a second calculation timestamp corresponding to the target second timestamp;

[0076] Determine an associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp;

[0077] The associated first detection point cluster is used as a first detection point cluster that has a temporal association relationship with the second detection point cloud corresponding to the target second timestamp.

[0078] In some possible embodiments, when the synchronization module determines the associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp, the synchronization module is configured as follows:

[0079] Performing linear interpolation processing on the detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster; or

[0080] Mean processing is performed on the coordinates of the first detection point in the detection point cluster corresponding to the first calculation timestamp and the first detection point in the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster.

[0081] In some possible embodiments, when the synchronization module performs spatial synchronization processing on the first detection point cluster and the second detection point cloud, it is configured as follows:

[0082] The coordinates of the first detection point are transformed based on the rotation parameters and the translation parameters, and the coordinates of the second detection point are transformed based on the rotation parameters and the translation parameters to obtain the coordinates of the first detection point and the coordinates of the second detection point after spatial synchronization; wherein the rotation parameters are used to rotate the coordinate axes in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point to the left to coincide with the coordinate axes of the spatial synchronization coordinate system; and the translation parameters are used to determine the coordinate components of the origin of the spatial synchronization coordinate system in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point.

[0083] In some possible embodiments, the rotation parameter and the translation parameter are obtained according to the following method:

[0084] Using a laser radar to measure a target object, obtain at least one measurement point of the target object and coordinates corresponding to the measurement point; and determine the mean value of the coordinates of the measurement point;

[0085] Measuring the target object using a millimeter wave radar to obtain a first coordinate of the target object;

[0086] A rotation parameter and a translation parameter are determined based on the mean of the coordinates and the first coordinate.

[0087] In some possible embodiments, the coordinates of the first detection point include radial distance, azimuth angle and elevation angle;

[0088] When the expansion module performs area expansion on the coordinates of each first detection point in the first detection point cluster after synchronization processing according to the standard deviation, it is configured as follows:

[0089] The following process is performed for each first detection point in the first detection point cluster:

[0090] Determine a radial distance standard deviation, an azimuth standard deviation, and an elevation standard deviation of the first detection point according to a signal-to-noise ratio of the first detection point;

[0091] Determine a first difference between the radial distance of the first detection point and the standard deviation of the radial distance, and determine a first sum between the radial distance of the first detection point and the standard deviation of the radial distance;

[0092] Determine an expansion area of ​​the radial distance according to the first difference and the first sum;

[0093] Determine a second difference between the azimuth of the first detection point and the azimuth standard deviation, and determine a second sum between the azimuth of the first detection point and the azimuth standard deviation;

[0094] Determine an extended area of ​​the azimuth according to the second difference and the second sum;

[0095] Determine a third difference between the pitch angle of the first detection point and the standard deviation of the pitch angle, and determine a third sum between the pitch angle of the first detection point and the standard deviation of the pitch angle;

[0096] Determining an extended area of ​​the pitch angle according to the third sum;

[0097] An expansion area of ​​the first detection point is determined according to the expansion area of ​​the radial distance, the expansion area of ​​the azimuth angle, and the expansion area of ​​the elevation angle.

[0098] In some possible embodiments, when the fusion module performs associating the second detection point that falls in the area corresponding to the first detection point with the first detection point, the fusion module is configured to:

[0099] If a plurality of second detection points fall within the extended area of ​​the first detection point, determining a second detection point with the largest signal-to-noise ratio in the extended area;

[0100] The second detection point with the largest signal-to-noise ratio is used as the second detection point associated with the first detection point.

[0101] In some possible embodiments, the first detection point cluster includes: a plurality of first detection points, the second detection point cloud includes: a plurality of second detection points, and the step of fusing information of the mutually related first detection points and second detection points to obtain fused detection points includes:

[0102] For each second detection point, perform the following process:

[0103] Using the physical parameter information of the first detection point associated with the second detection point as the physical parameter information of the fused detection point corresponding to the second detection point;

[0104] Using the spatially synchronized coordinates of the second detection point as the coordinates of the fused detection point corresponding to the second detection point;

[0105] The fused detection point is formed based on the physical parameter information of the fused detection point and the coordinates of the fused detection point.

[0106] In some possible embodiments, when the heading angle determination module determines the heading angle of each vehicle based on the fused detection point cloud, it is configured as follows:

[0107] For each frame of each vehicle, the fused detection point cloud is executed:

[0108] Determining an initial heading angle of the vehicle based on the fused detection point cloud;

[0109] Filter out a target fusion detection point according to the signal-to-noise ratio of each fusion detection point in the fusion detection point cloud;

[0110] Determine a wheel position difference of the vehicle based on the fused detection point cloud and a fused detection point cloud of a previous frame;

[0111] The initial heading angle, the target fusion detection point and the wheel position difference are input into a pre-trained motion model to obtain the heading angle of the vehicle.

[0112] In the third aspect, another embodiment of the present application also provides an electronic device, comprising at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any method provided in the embodiment of the first aspect of the present application.

[0113] In a fourth aspect, another embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is used to enable a computer to execute any method provided in the embodiment of the first aspect of the present application.

[0114] Other features and advantages of the present application will be described in the following description, and partly become apparent from the description, or be understood by practicing the present application. The purpose and other advantages of the present application can be realized and obtained by the structures specifically pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. Obviously, the drawings introduced below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0116] Figure 1 An application scenario diagram of a method for determining a heading angle provided in an embodiment of the present application;

[0117] Figure 2 A schematic diagram of the overall process of a method for determining a heading angle provided in an embodiment of the present application;

[0118] Figure 3 A schematic diagram of a timestamp marking method for determining a heading angle provided in an embodiment of the present application;

[0119] Figure 4 A schematic diagram of determining an associated first detection point cluster in a method for determining a heading angle provided in an embodiment of the present application;

[0120] Figure 5 A schematic diagram of marking a third timestamp of a method for determining a heading angle provided in an embodiment of the present application;

[0121] Figure 6 A schematic diagram of determining an associated first detection point cluster in a method for determining a heading angle provided in an embodiment of the present application;

[0122] Figure 7 A schematic diagram of parameter determination of a method for determining a heading angle provided in an embodiment of the present application;

[0123] Fig. 8A A schematic diagram of a region expansion of a method for determining a heading angle provided in an embodiment of the present application;

[0124] Figure 8B A schematic diagram of an extended area of ​​a heading angle determination method provided in an embodiment of the present application;

[0125] Fig. 9 A schematic diagram of information fusion of a method for determining a heading angle provided in an embodiment of the present application;

[0126] Fig.10A schematic diagram of determining a heading angle according to a method for determining a heading angle provided in an embodiment of the present application;

[0127] Fig.11 A schematic diagram of determining an initial heading angle in a heading angle determination method provided in an embodiment of the present application;

[0128] Fig.12 A schematic diagram of the overall process of a method for determining a heading angle provided in an embodiment of the present application;

[0129] Fig.13 A schematic diagram of a device for determining a heading angle provided in an embodiment of the present application;

[0130] Fig.14 A schematic diagram of an electronic device for a method for determining a heading angle provided in an embodiment of the present application. DETAILED DESCRIPTION

[0131] In order to enable ordinary persons in the art to better understand the technical solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0132] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application 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 interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0133] The inventor has found that the heading angle of a moving vehicle is a very important parameter when estimating the moving state of a moving vehicle at the next moment.

[0134] In the related technology, millimeter-wave radar or laser radar is mostly used to estimate the heading angle of a moving vehicle. When using millimeter-wave radar to estimate the heading angle of a moving vehicle, the millimeter-wave radar is first used to measure the position of the vehicle and its Doppler information, and then the corresponding heading angle estimation model is established in combination with the selected tracking model. However, the measurement accuracy of this method is limited by the error of the millimeter-wave radar in measuring the position of the moving vehicle, the error of the Doppler measurement, and the error of the heading angle estimation model; when using laser radar to estimate the heading angle of a moving vehicle, the point cloud obtained by detection is first segmented, and then the shape of the point cloud is estimated based on the point cloud to obtain the heading angle of the moving vehicle. However, the accuracy of the heading angle measurement of this method is limited by the accuracy of the point cloud obtained by detection.

[0135] In view of this, the present application proposes a heading angle determination method, device, electronic device and storage medium for solving the above problems. The inventive concept of the present application can be summarized as follows: firstly, a millimeter wave radar and a laser radar are respectively used to continuously detect the target scene to obtain a continuous multi-frame first detection point cluster and a continuous multi-frame second detection point cloud; then the first detection point cluster and the second detection point cloud are subjected to time synchronization processing and space synchronization processing; and the coordinates of each first detection point in the first detection point cluster after synchronization processing are regionally expanded according to the standard deviation to obtain the extended area of ​​the first detection point, and the second detection point falling in the extended area is determined according to the coordinates of the second detection point; wherein the standard deviation is the standard deviation of the physical parameter associated with the first detection point; then the second detection point falling in the extended area is associated with the first detection point, and the information of the mutually associated first detection point and second detection point is fused to obtain a fused detection point; the fused detection points constitute a fused detection point cloud; finally, the fused detection point cloud composed of the fused detection points is segmented to obtain a fused detection point cloud corresponding to each vehicle in the target scene, and the heading angle of each vehicle is determined based on the fused detection point cloud.

[0136] In order to facilitate understanding of a method for determining a heading angle provided in an embodiment of the present application, the following is a detailed description with reference to the accompanying drawings:

[0137] like Figure 1 The figure shows an application scenario diagram of a heading angle determination method in an embodiment of the present application. The figure includes: a laser radar 10, a millimeter wave radar 20, a server 30, and a memory 40; wherein: the millimeter wave radar 20 is used to continuously detect the target scene to obtain a continuous multi-frame first detection point cluster, and the laser radar 10 is used to continuously detect the target scene to obtain a continuous multi-frame second detection point cloud; wherein the target scene includes at least one vehicle; the laser radar and the millimeter wave radar report the first detection point cluster and the second detection point cloud obtained by themselves to the server respectively; the server 30 performs time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain a first detection point cluster and a second detection point cloud that are associated in time and space; then according to the standard The coordinates of each first detection point in the first detection point cluster after synchronization processing are expanded to obtain the extended area of ​​the first detection point, and the second detection point falling in the extended area is determined according to the coordinates of the second detection point; wherein the standard deviation is the standard deviation of the physical parameter associated with the first detection point; the second detection point falling in the extended area is associated with the first detection point, and the information of the mutually associated first detection point and second detection point is fused to obtain a fused detection point; the fused detection points constitute a fused detection point cloud; the fused detection point cloud constituted by the fused detection points is segmented to obtain a fused detection point cloud corresponding to each vehicle in the target scene, and the heading angle of each vehicle is determined based on the fused detection point cloud.

[0138] The description in this application only details a single server or memory, but those skilled in the art should understand that the laser radar 10, millimeter wave radar 20, server 30, and memory 40 shown are intended to represent the operation of the laser radar 10, millimeter wave radar 20, server 30, and memory 40 involved in the technical solution of this application. The detailed description of a single laser radar 10, millimeter wave radar 20, server 30, and memory 40 is at least for the convenience of explanation, and does not imply any limitation on the number, type, or location of the laser radar 10, millimeter wave radar 20, server 30, and memory 40. It should be noted that if Figure 1 The addition of additional modules or removal of individual modules from the depicted environment does not change the underlying concepts of the example embodiments of the present application.

[0139] It should be noted that the memory in the embodiment of the present application may be, for example, a cache system, a hard disk storage, a memory storage, etc. In addition, the heading angle determination method proposed in the present application is not only applicable to Figure 1 The application scenario shown is also applicable to any device that needs to determine the heading angle.

[0140] like Figure 2 FIG. 1 is a flow chart of a method for determining a heading angle provided in an embodiment of the present application, wherein:

[0141] In step 201: a millimeter-wave radar is used to continuously detect the target scene to obtain a first detection point cluster of multiple consecutive frames, and a laser radar is used to continuously detect the target scene to obtain a second detection point cloud of multiple consecutive frames; wherein the target scene includes at least one vehicle;

[0142] In step 202: performing time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain the first detection point cluster and the second detection point cloud that are associated in both time and space;

[0143] In step 203: the coordinates of each first detection point in the first detection point cluster after synchronization processing are expanded according to the standard deviation to obtain an expanded area of ​​the first detection point, and the second detection point falling in the expanded area is determined according to the coordinates of the second detection point; wherein the standard deviation is the standard deviation of the physical parameter associated with the first detection point;

[0144] In step 204: the second detection point falling in the extended area is associated with the first detection point, and the information of the first detection point and the second detection point associated with each other is fused to obtain a fused detection point; the fused detection point constitutes a fused detection point cloud;

[0145] In step 205: the fused detection point cloud composed of the fused detection points is segmented to obtain the fused detection point cloud corresponding to each vehicle in the target scene, and the heading angle of each vehicle is determined based on the fused detection point cloud.

[0146] For ease of understanding, the following are Figure 2 The steps shown are explained in detail:

[0147] First, the information of the first detection point cluster and the second detection point cloud is described, wherein: the high-performance millimeter-wave radar continuously detects the target scene, and the continuous multi-frame first detection point cluster obtained, each frame of the first detection point cluster includes: the first detection point, the radial distance of the first detection point, Doppler velocity information, azimuth, pitch angle information, radar cross-sectional area (RadarCross Section, RCS), point quality, first timestamp, signal-to-noise ratio and other information; the laser radar continuously detects the target scene, and the continuous multi-frame second detection point cloud obtained, each frame of the second detection point cloud includes: the second detection point, the second timestamp.

[0148] The following describes in detail a method for determining a heading angle provided in an embodiment of the present application in combination with the first detection point cluster and the second detection point cloud:

[0149] 1. Time synchronization

[0150] In the present application, in order to ensure that the final heading angle of the moving vehicle is more accurate, it is necessary to time synchronize the first detection point cluster and the second detection point cloud. In the present application, it can be implemented as follows: for each second timestamp, determine the first calculation timestamp and the second calculation timestamp corresponding to the second timestamp, and based on the first calculation timestamp, the second calculation timestamp and the second timestamp, determine the associated first detection point cluster that has an associated relationship with the second detection point cloud corresponding to the second timestamp.

[0151] Among them: the first timestamp is obtained by marking the first detection point cluster in each frame obtained by the millimeter wave radar continuously detecting the target scene according to the first preset frequency, and the second timestamp is obtained by marking the second detection point cloud in each frame obtained by the laser radar continuously detecting the target scene according to the second preset frequency; the first calculation timestamp is the first timestamp that is located before the second timestamp in time sequence and has the smallest time interval with the second timestamp, and the second calculation timestamp is the first timestamp that is located after the second timestamp in time sequence and has the smallest time interval with the second timestamp.

[0152] For example, if the frequency of the millimeter wave radar is 20 Hz and the frequency of the laser radar is 12 Hz, then the first timestamp and the second timestamp are as follows: Figure 3 As shown, the first calculation timestamp of timestamp t1 is a0, and the second calculation timestamp is a1.

[0153] Determining an associated first detection point cluster having an associated relationship with a second detection point cloud corresponding to the second timestamp based on the first calculation timestamp, the second calculation timestamp and the second timestamp can be specifically implemented as follows: Figure 4 Steps shown:

[0154] In step 401: obtaining a third timestamp, wherein the third timestamp is determined to be obtained by marking with a first preset frequency starting from a first time interval after the laser radar marks the first second timestamp; wherein the first time interval is obtained according to the second preset frequency;

[0155] For example, assuming that the frequency of the laser radar is 12 Hz and the frequency of the millimeter-wave radar is 20 Hz, the frequency of the third timestamp is 12 Hz. Since the frequency of the millimeter-wave radar is 20 Hz, the first time interval is the inverse of the millimeter-wave radar frequency, which is 50 milliseconds. Therefore, the third timestamp is Figure 5 shown.

[0156] In step 402: for each third timestamp, determine a target second timestamp that has the smallest timing difference with the third timestamp and is before the third timestamp;

[0157] For example: Continue with Figure 5 For example, taking c0 as an example, the target second timestamp is b0.

[0158] In step 403: obtaining a first calculation timestamp and a second calculation timestamp corresponding to the target second timestamp;

[0159] by Figure 5 For example, the first calculation timestamp corresponding to the target second timestamp b0 is a0, and the second calculation timestamp is a1.

[0160] In step 404: determining an associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp;

[0161] In the embodiments of the present application, the following two methods can be specifically implemented:

[0162] 1) Performing linear interpolation processing on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp to obtain an associated first detection point cluster.

[0163] For example, the first detection point cluster corresponding to the first calculation timestamp is point cluster A, and the first detection point cluster corresponding to the second calculation timestamp is point cluster B. Figure 6As shown, point cluster A and point cluster B are associated according to the coordinates of the midpoints of point cluster A and point cluster B, that is, the two points with the closest coordinates are used as the two associated points, for example, 1 is associated with point 1', 2 is associated with point 2', and so on and so forth. Then linear interpolation is performed on the associated points to obtain interpolation points, and the associated first detection point cluster is obtained based on the interpolation points.

[0164] 2) Performing mean processing on the coordinates of the first detection point in the detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp to obtain an associated first detection point cluster.

[0165] For example: Continue with Figure 6 For example, firstly, point cluster A and point cluster B are associated according to the coordinates of the midpoints of point cluster A and point cluster B, that is, the two points with the closest coordinates are used as the two associated points, for example, 1 is associated with point 1', 2 is associated with point 2', and so on and so forth. Then, the two associated points are averaged to obtain the associated first detection point cluster.

[0166] In step 405: the associated first detection point cluster is used as the first detection point cluster that has a temporal association relationship with the second detection point cloud corresponding to the target second timestamp.

[0167] 2. Space synchronization

[0168] In an embodiment of the present application, in order to facilitate the subsequent calculation of the heading angle, it is necessary to perform spatial synchronization processing on the first detection point cluster and the second detection point cloud, which can be specifically implemented as follows: coordinate transformation of the first detection point based on rotation parameters and translation parameters, and coordinate transformation of the second detection point based on rotation parameters and translation parameters, to obtain the coordinates of the first detection point and the second detection point after spatial synchronization; wherein the rotation parameters are used to rotate each coordinate axis in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point to the left to coincide with the coordinate axis of the spatial synchronization coordinate system; the translation parameters are used to determine the coordinate components of the origin of the spatial synchronization coordinate system in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point. In the present application, the translation parameters and the rotation parameters can be used as follows: Figure 7 The steps shown determine where:

[0169] In step 701: a laser radar is used to measure a target object to obtain at least one measurement point of the target object and coordinates corresponding to the measurement point; and a mean value of the coordinates of the measurement point is determined;

[0170] In step 702: using a millimeter wave radar to measure the target object to obtain a first coordinate of the target object;

[0171] In step 703: a rotation parameter and a translation parameter are determined based on the mean value of the coordinates and the first coordinate.

[0172] Of course, it should be noted that, in specific implementation, the present application does not limit the execution order of step 701 and step 702, that is, step 701 may be executed first and then step 702, or step 702 may be executed first and then step 701, or step 701 and step 702 may be executed simultaneously.

[0173] For example: select a small metal panel (target object) and place it in front of the millimeter wave radar, and meet the millimeter wave radar far-field measurement conditions. At this time, the millimeter wave can accurately measure the position of the metal plate, with a horizontal accuracy of 0.1 degrees and a pitch accuracy of 0.2 degrees. At the same time, the laser radar can accurately measure the spatial position of the stationary metal plate. The laser radar has a high azimuth measurement resolution, and the metal plate will get multiple measurement points. At this time, the average value L of the coordinates of these points can be taken to determine the translation parameters and rotation parameters with the millimeter wave radar measurement value Rm and formula 1.

[0174] L*R+T=Rm, (Formula 1)

[0175] Where: L is the average value of the laser radar measurement point, Rm is the coordinate value of the millimeter wave radar measurement point, R is the rotation parameter, and T is the translation parameter.

[0176] In summary, by moving the position of the metal plate, multiple sets of L and Rm values ​​can be obtained, and then the rotation parameters and translation parameters can be obtained.

[0177] 3. Regional expansion

[0178] In the present application, when the region expansion is performed on each first detection point in the first detection point cluster, it can be implemented as follows: Fig. 8A Steps shown:

[0179] In step 801: determining the radial distance standard deviation, azimuth standard deviation and elevation standard deviation of the first detection point according to the signal-to-noise ratio of the first detection point;

[0180] In step 802: determining a first difference between the radial distance of the first detection point and the standard deviation of the radial distance, and determining a first sum between the radial distance of the first detection point and the standard deviation of the radial distance;

[0181] In step 803: determining an expansion area of ​​the radial distance according to the first difference and the first sum;

[0182] In step 804: determining a second difference between the azimuth of the first detection point and the azimuth standard deviation, and determining a second sum between the azimuth of the first detection point and the azimuth standard deviation;

[0183] In step 805: determining an extended area of ​​the azimuth angle according to the second difference and the second sum;

[0184] In step 806: determining a third difference between the pitch angle of the first detection point and the standard deviation of the pitch angle, and determining a third sum between the pitch angle of the first detection point and the standard deviation of the pitch angle;

[0185] In step 807: determining an extended area of ​​the pitch angle according to the third sum value;

[0186] In step 808: the expansion area of ​​the first detection point is determined according to the expansion area of ​​the radial distance, the expansion area of ​​the azimuth angle and the expansion area of ​​the pitch angle.

[0187] It should be noted that the present application does not limit the order of determining the extended area of ​​the radial distance, determining the extended area of ​​the azimuth angle, and determining the extended area of ​​the elevation angle. The technicians can set the execution order according to the needs. Fig. 8A This is given as just one example.

[0188] For example: the coordinates of the first measurement point A are (r, a, e), where r represents the radial distance, a represents the azimuth, and e represents the elevation. At this time, the extended area of ​​the radial distance is (r-Δr, r+Δr), the extended area of ​​the azimuth is (a-Δa, a+Δa), and the extended area of ​​the elevation is (e-Δe, e+Δe), where Δr is the standard deviation of r, Δa is the standard deviation of a, and Δe is the standard deviation of e. Δ is the standard deviation of r, Δ is the standard deviation of a, and Δ is the standard deviation of e. The extended area of ​​point A is as follows: Figure 8B shown.

[0189] When determining the standard deviation, a radar target simulator can be used for testing. For example, when measuring the standard deviation, set the target to a fixed distance, then adjust the target reflection power, and perform multiple measurements each time the adjustment is made to calculate the sample standard deviation. When calculating the azimuth standard deviation, set the reflection power at different azimuths according to the antenna pattern, and then perform multiple measurements each time the adjustment is made to calculate the sample standard deviation of the azimuth.

[0190] In summary, the expansion area of ​​the first detection point can be determined according to the expansion area of ​​the radial distance, the expansion area of ​​the azimuth angle, and the expansion area of ​​the elevation angle.

[0191] 4. Relationship

[0192] In the present application, when a second detection point falling in the corresponding area of ​​the first detection point is associated with the first detection point, if multiple second detection points fall in the extended area of ​​the first detection point, the second detection point with the largest signal-to-noise ratio in the extended area is determined; and the second detection point with the largest signal-to-noise ratio is used as the second detection point associated with the first detection point.

[0193] For example, the first detection point is (4, 5, 10) and the extended area of ​​the first detection point is [(4-0.2, 4+0.2), (5-0.3, 5+0.3), (10-0.1, 10+0.1)], then determine the second detection point in the extended area. If there are detection points A, B, C, and D in the extended area, where the signal-to-noise ratio of point A is 124, the signal-to-noise ratio of point B is 106, the signal-to-noise ratio of point C is 114, and the signal-to-noise ratio of point D is 103, then according to the signal-to-noise ratios of detection points A, B, C, and D, point A with the largest signal-to-noise ratio is used as the second detection point associated with the first detection point.

[0194] 5. Fusion

[0195] In the present application, when information fusion is performed on the mutually related first detection points and second detection points, the following is performed for each second detection point: Fig. 9 The steps shown, where:

[0196] In step 901: the physical parameter information of the first detection point associated with the second detection point is used as the physical parameter information of the fused detection point corresponding to the second detection point;

[0197] In step 902: the spatially synchronized coordinates of the second detection point are used as the coordinates of the fused detection point corresponding to the second detection point;

[0198] In step 903: a fused detection point is formed based on the physical parameter information of the fused detection point and the coordinates of the fused detection point.

[0199] For example: the first detection point A and the second detection point B are associated detection points. The physical parameter information of the first detection point A, namely the Doppler velocity information, RCS, point quality, first timestamp, and signal-to-noise ratio, is used as the physical parameter information of the fused detection point C, and the coordinates of the second detection point B are used as the coordinates of the fused detection point C.

[0200] 6. Determine the heading angle

[0201] In the present application, since there may be more than one moving vehicle appearing in the target scene, it is necessary to segment the fused detection point cloud before determining the initial heading angle of the vehicle based on the fused detection point cloud, and a pre-trained segmentation model is used when segmenting the fused detection point cloud. The segmentation model in the present application can be a Transformer model or a deep network model (Deep Neural Networks, DNN). When training the segmentation model, a conventional network model training method can be used, for example: first obtain a point cloud containing multiple vehicles and the segmentation results of the point cloud, use the point cloud as the input of the DNN, use the segmentation results as the expected output to train the DD model, and adjust the parameters of the DNN model according to the gap between the output result and the segmentation result until the training converges. This application does not limit the method of training the network model, and the technicians can set the training method according to their needs.

[0202] In the present application, the heading angle of each vehicle can be determined based on the fusion detection points as follows: Fig.10 The steps shown, where:

[0203] In step 1001: determining the initial heading angle of the vehicle based on the fused detection point cloud;

[0204] In the present application, in order to avoid the waste of computing power caused by all fused detection point clouds entering subsequent calculations, it is necessary to filter the fused detection point cloud before determining the initial heading angle of the vehicle to remove the static point cloud in the fused detection point cloud. Specifically, it can be implemented as follows: determine the value of the absolute Doppler velocity of each fused detection point in the fused detection point cloud to determine whether the fused detection point cloud is moving. For example, if there is a fused detection point whose absolute Doppler value is not zero, then the fused detection point cloud is considered to be moving; if there is no fused detection point whose absolute Doppler value is not zero in the fused detection point cloud, then the fused detection point cloud is considered to be a static point cloud and the fused detection point cloud is removed.

[0205] In addition to determining the motion state of the fused detection point cloud based on the absolute Doppler value of the fused detection point, it is also possible to determine whether the point cloud is moving based on the position information of the fused detection points in the fused detection point clouds of the previous and next frames. Specifically, it can be implemented as follows: determine the position information of each fused detection point in the fused detection point cloud of the current frame, and determine the position information of each fused detection point in the fused detection point cloud of the previous frame. If there are fused detection points whose positions have changed, it can be considered that the fused detection point cloud is moving.

[0206] When determining the initial heading angle for the filtered fused detection point cloud, the shape of each segmented fused detection point cloud can be estimated, for example: take a plane with the same height, and then use the Search RectangleBoard Filter (SRBF) method; the heading angle azimuth value is between -180 degrees and 180 degrees, and a coarse search can be used first, and then a fine search can be used. The coarse search angle interval is 2 degrees, and the fine search is 0.2 degrees. The coarse search and fine search processing methods are the same, and the coarse search is used as an example for explanation below:

[0207] like Fig.11 As shown in the figure, the heading angle azimuth value is selected as 0 degrees. At this time, the X-axis is the direction indicated by the angle, and the Y-axis is the direction perpendicular to it, satisfying the right-hand rule. Then the fused detection point cloud is projected to the coordinate, and the maximum value of the horizontal coordinate and the maximum value of the vertical coordinate are calculated to fit into a rectangle. For each fused detection point in the fused detection point cloud, the distance from the point to the boundaries of the fitted rectangle is calculated, and the distance is a positive value. Select the smallest one, and then add up the minimum distances of all points to get the cumulative minimum distance at the angle.

[0208] For example: Fig.11 Taking point A in as an example, the distances to the four boundaries of point A are l1, l2, l3, and l4 respectively; among which l1 is the minimum distance, then l1 is selected as the minimum distance of point A.

[0209] The above operation is performed for all angles, and then the angle that minimizes the cumulative sum of the minimum distances is selected and used as the initial heading angle of the vehicle. The process of the fine search is the same as above and will not be repeated here.

[0210] In the specific implementation, the length or width of the rectangle may be small. Therefore, when the length or width of the rectangle is less than a certain threshold, it can be considered that the measured is an edge. Then, when calculating the heading angle, the heading angle of the previous frame point cloud can be used as the heading angle of the current frame fused detection point cloud.

[0211] In step 1002: select a target fusion detection point according to the signal-to-noise ratio of each fusion detection point in the fusion detection point cloud;

[0212] In this application, in order to ensure the accuracy of the determined heading angle, the target fusion detection point is selected based on the signal-to-noise ratio, which can be implemented as follows: the four fusion detection points with the largest signal-to-noise ratio are selected in sequence as the target fusion detection points. A signal-to-noise ratio threshold can also be set, and the fusion detection points with a signal-to-noise ratio greater than the threshold are used as fusion detection points. This application does not limit this.

[0213] In step 1003: determining the wheel position difference of the vehicle based on the fused detection point cloud and the fused detection point cloud of the previous frame;

[0214] When determining the wheel position difference of the vehicle, the fused detection point cloud can be clustered and analyzed, the outline of the vehicle can be obtained based on the fused detection point cloud, and the measured wheel position can be obtained based on the outline of the vehicle. Then, the fused detection point cloud of the previous frame is processed in the same way to obtain the wheel position of the fused detection point cloud of the previous frame, and then the wheel position difference of the vehicle can be obtained.

[0215] The wheel position in the fused detection point cloud can also be determined based on the DNN network model, and then combined with the wheel position in the fused detection point cloud of the previous frame, the wheel position difference between the previous and next moments can be obtained. The wheel position can also be determined by referring to the common position difference between the wheel and the vehicle body; the wheel position can also be determined by analyzing the micro-Doppler distribution state at different positions in combination with micro-Doppler, and then the wheel position difference can be obtained. This application does not limit the method for obtaining the wheel position difference. All methods that can obtain the wheel position difference are applicable to this application, and technicians can choose according to their needs.

[0216] In the present application, in order to further make the determined heading angle more accurate, the wheel position difference is the position difference of the driving wheels, that is: if the vehicle is a rear-wheel drive vehicle, the wheel position difference is the wheel position difference of the rear wheels; if the vehicle is a front-wheel drive vehicle, the wheel position difference is the wheel position difference of the front wheels; if the vehicle is a four-wheel drive vehicle, the wheel position difference is the wheel position difference of the four wheels.

[0217] In step 1004: the initial heading angle, the target fusion detection point and the wheel position difference are input into a pre-trained motion model to obtain the heading angle of the vehicle.

[0218] The following is an example of inputting four target fusion detection points and the wheel position difference of the rear wheels of a rear-wheel drive vehicle:

[0219] The motion model can be the Constant Turn Rate and Velocity (CTRV), and the filtering method can be the Extended Kalman Filter (EKF). The output of the motion model is [V, yawrate, Orientation], where V represents the vehicle speed, yawrate represents the vehicle yaw rate, and Orientation represents the vehicle heading angle. The input is [Orientation_Mea, VDoppler1, VDoppler2, VDoppler3, VDoppler4, ΔΔ, ΔΔo, where Orientation_Mea is the initial heading angle, VDoppler1, VDoppler2, VDoppler3, and VDoppler4 are the Doppler information of the target fusion point, and ΔΔ and ΔΔ are the wheel position differences. When the extended Kalman filter is used for calculation, the measurement variance and system variance are also required. The measurement variance of Orientation_Mea can be obtained by accumulating the strength of the fused detection points in the fused detection point cloud and the sample variance when the fused detection point cloud is estimated; the system variance of VDoppler is determined according to the signal-to-noise ratio of the fused detection points in the fused detection point cloud. The system variance can also be obtained by combining the deviation between the output result of the motion model at the previous moment and the actual heading angle of the vehicle and the deviation obtained at the previous moment. For example, the difference between the deviations of the previous and next moments can be calculated, and the system variance can be determined based on the relationship between the difference and the threshold.

[0220] For ease of understanding, the overall process of a heading angle estimation method provided in an embodiment of the present application is described in detail below. Fig.12 As shown, where:

[0221] In step 1201: a millimeter-wave radar is used to continuously detect the target scene to obtain a first detection point cluster of multiple consecutive frames, and a laser radar is used to continuously detect the target scene to obtain a second detection point cloud of multiple consecutive frames;

[0222] In step 1202: performing time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain the first detection point cluster and the second detection point cloud that are associated in both time and space;

[0223] In step 1203: the coordinates of each first detection point in the first detection point cluster after synchronization processing are expanded according to the standard deviation to obtain an expanded area of ​​the first detection point, and the second detection point falling in the expanded area is determined according to the coordinates of the second detection point;

[0224] In step 1204: the second detection point falling in the extended area is associated with the first detection point, and the information of the first detection point and the second detection point associated with each other is fused to obtain a fused detection point; the fused detection point constitutes a fused detection point cloud;

[0225] In step 1205: segmenting the fused detection point cloud formed by the fused detection points to obtain the fused detection point cloud corresponding to each vehicle in the target scene;

[0226] In step 1206: determining the initial heading angle of the vehicle based on the fused detection point cloud;

[0227] In step 1207: select the target fusion detection point according to the signal-to-noise ratio of each fusion detection point in the fusion detection point cloud;

[0228] In step 1208: determining the wheel position difference of the vehicle based on the fused detection point cloud and the fused detection point cloud of the previous frame;

[0229] In step 1209: the initial heading angle, the target fusion detection point and the wheel position difference are input into the pre-trained motion model to obtain the heading angle of the vehicle.

[0230] like Fig.13 As shown, based on the same inventive concept, a heading angle determination device 1300 is proposed, comprising:

[0231] The detection module 13001 is used to continuously detect the target scene using a millimeter wave radar to obtain a first detection point cluster of multiple consecutive frames, and to continuously detect the target scene using a laser radar to obtain a second detection point cloud of multiple consecutive frames; wherein the target scene includes at least one vehicle;

[0232] A synchronization module 13002 is used to perform time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain a first detection point cluster and a second detection point cloud that are associated in both time and space;

[0233] An expansion module 13003 is used to perform area expansion on the coordinates of each first detection point in the first detection point cluster after synchronization processing according to a standard deviation to obtain an extended area of ​​the first detection point, and determine a second detection point falling in the extended area according to the coordinates of the second detection point; wherein the standard deviation is a standard deviation of a physical parameter associated with the first detection point;

[0234] A fusion module 13004 is used to associate the second detection point falling in the extended area with the first detection point, and to fuse the information of the first detection point and the second detection point associated with each other to obtain a fused detection point; the fused detection point constitutes a fused detection point cloud;

[0235] The heading angle determination module 13005 is used to segment the fused detection point cloud composed of the fused detection points to obtain the fused detection point cloud corresponding to each vehicle in the target scene, and determine the heading angle of each vehicle based on the fused detection point cloud.

[0236] In some possible embodiments, when the synchronization module 13002 performs time synchronization processing on the first detection point cluster and the second detection point cloud, it is configured as follows:

[0237] For each second timestamp, determine a first calculation timestamp and a second calculation timestamp corresponding to the second timestamp; wherein the first timestamp is obtained by marking a first detection point cluster of each frame obtained by continuously detecting the target scene by the millimeter wave radar at a first preset frequency, and the second timestamp is obtained by marking a second detection point cloud of each frame obtained by continuously detecting the target scene by the laser radar at a second preset frequency; wherein the first calculation timestamp is a first timestamp that is located before the second timestamp in time sequence and has the smallest time interval with the second timestamp, and the second calculation timestamp is a first timestamp that is located after the second timestamp in time sequence and has the smallest time interval with the second timestamp;

[0238] An associated first detection point cluster having an association relationship with a second detection point cloud corresponding to the second timestamp is determined based on the first calculation timestamp, the second calculation timestamp, and the second timestamp.

[0239] In some possible embodiments, when the synchronization module 13002 determines the associated first detection point cluster having an association relationship with the second detection point cloud corresponding to the second timestamp based on the first calculation timestamp, the second calculation timestamp and the second timestamp, the synchronization module 13002 is configured as follows:

[0240] Obtain a third timestamp, wherein the third timestamp is determined to be obtained by marking at a second preset frequency starting from a first time interval after the laser radar marks the first second timestamp; wherein the first time interval is obtained according to the second preset frequency;

[0241] For each third timestamp, determine a target second timestamp that has the smallest timing difference with the third timestamp and is before the third timestamp;

[0242] Obtain a first calculation timestamp and a second calculation timestamp corresponding to the target second timestamp;

[0243] Determine an associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp;

[0244] The associated first detection point cluster is used as a first detection point cluster that has a temporal association relationship with the second detection point cloud corresponding to the target second timestamp.

[0245] In some possible embodiments, when the synchronization module 13002 determines the associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp, it is configured as follows:

[0246] Performing linear interpolation processing on the detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster; or

[0247] Mean processing is performed on the coordinates of the first detection point in the detection point cluster corresponding to the first calculation timestamp and the first detection point in the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster.

[0248] In some possible embodiments, when the synchronization module 13002 performs spatial synchronization processing on the first detection point cluster and the second detection point cloud, it is configured as follows:

[0249] The coordinates of the first detection point are transformed based on the rotation parameters and the translation parameters, and the coordinates of the second detection point are transformed based on the rotation parameters and the translation parameters to obtain the coordinates of the first detection point and the coordinates of the second detection point after spatial synchronization; wherein the rotation parameters are used to rotate the coordinate axes in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point to the left to coincide with the coordinate axes of the spatial synchronization coordinate system; and the translation parameters are used to determine the coordinate components of the origin of the spatial synchronization coordinate system in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point.

[0250] In some possible embodiments, the rotation parameter and the translation parameter are obtained according to the following method:

[0251] Using a laser radar to measure a target object, obtain at least one measurement point of the target object and coordinates corresponding to the measurement point; and determine the mean value of the coordinates of the measurement point;

[0252] Measuring the target object using a millimeter wave radar to obtain a first coordinate of the target object;

[0253] A rotation parameter and a translation parameter are determined based on the mean of the coordinates and the first coordinate.

[0254] In some possible embodiments, the coordinates of the first detection point include radial distance, azimuth angle and elevation angle;

[0255] When the expansion module 13003 performs area expansion on the coordinates of each first detection point in the first detection point cluster after synchronization processing according to the standard deviation, it is configured as follows:

[0256] The following process is performed for each first detection point in the first detection point cluster:

[0257] Determine a radial distance standard deviation, an azimuth standard deviation, and an elevation standard deviation of the first detection point according to a signal-to-noise ratio of the first detection point;

[0258] Determine a first difference between the radial distance of the first detection point and the standard deviation of the radial distance, and determine a first sum between the radial distance of the first detection point and the standard deviation of the radial distance;

[0259] Determine an expansion area of ​​the radial distance according to the first difference and the first sum;

[0260] Determine a second difference between the azimuth of the first detection point and the azimuth standard deviation, and determine a second sum between the azimuth of the first detection point and the azimuth standard deviation;

[0261] Determine an extended area of ​​the azimuth according to the second difference and the second sum;

[0262] Determine a third difference between the pitch angle of the first detection point and the standard deviation of the pitch angle, and determine a third sum between the pitch angle of the first detection point and the standard deviation of the pitch angle;

[0263] Determining an extended area of ​​the pitch angle according to the third sum;

[0264] An expansion area of ​​the first detection point is determined according to the expansion area of ​​the radial distance, the expansion area of ​​the azimuth angle, and the expansion area of ​​the elevation angle.

[0265] In some possible embodiments, when the fusion module 13004 performs associating the second detection point that falls in the area corresponding to the first detection point with the first detection point, the fusion module 13004 is configured to:

[0266] If a plurality of second detection points fall within the extended area of ​​the first detection point, determining a second detection point with the largest signal-to-noise ratio in the extended area;

[0267] The second detection point with the largest signal-to-noise ratio is used as the second detection point associated with the first detection point.

[0268] In some possible embodiments, the first detection point cluster includes: a plurality of first detection points, the second detection point cloud includes: a plurality of second detection points, and the fusion module 13004 performs information fusion of the mutually related first detection points and second detection points to obtain fused detection points, including:

[0269] For each second detection point, perform the following process:

[0270] Using the physical parameter information of the first detection point associated with the second detection point as the physical parameter information of the fused detection point corresponding to the second detection point;

[0271] Using the spatially synchronized coordinates of the second detection point as the coordinates of the fused detection point corresponding to the second detection point;

[0272] The fused detection point is formed based on the physical parameter information of the fused detection point and the coordinates of the fused detection point.

[0273] In some possible embodiments, when the heading angle determination module 13005 determines the heading angle of each vehicle based on the fused detection point cloud, it is configured as follows:

[0274] For each frame of each vehicle, perform the fused detection point cloud:

[0275] Determining an initial heading angle of the vehicle based on the fused detection point cloud;

[0276] Filter out a target fusion detection point according to the signal-to-noise ratio of each fusion detection point in the fusion detection point cloud;

[0277] Determining a wheel position difference of the vehicle based on the fused detection point cloud and a fused detection point cloud of a previous frame;

[0278] The initial heading angle, the target fusion detection point and the wheel position difference are input into a pre-trained motion model to obtain the heading angle of the vehicle.

[0279] After introducing the heading angle determination method and device according to the exemplary embodiment of the present application, next, an electronic device according to another exemplary embodiment of the present application is introduced.

[0280] Those skilled in the art will appreciate that various aspects of the present application may be implemented as a system, method or program product. Therefore, various aspects of the present application may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to as "circuit", "module" or "system" herein.

[0281] In some possible implementations, the electronic device according to the present application may include at least one processor and at least one memory, wherein the memory stores program code, and when the program code is executed by the processor, the processor executes the steps in the heading angle determination method according to various exemplary implementations of the present application described above in this specification.

[0282] Refer to the following Fig.14 The electronic device 130 according to this embodiment of the present application is described. Fig.14 The electronic device 130 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0283] like Fig.14 As shown, the electronic device 130 is in the form of a general electronic device. The components of the electronic device 130 may include but are not limited to: the at least one processor 131, the at least one memory 132, and a bus 133 connecting different system components (including the memory 132 and the processor 131).

[0284] Bus 133 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a processor, or a local bus using any of a variety of bus architectures.

[0285] The memory 132 may include a readable medium in the form of a volatile memory, such as a random access memory (RAM) 1321 and / or a cache memory 1322 , and may further include a read-only memory (ROM) 1323 .

[0286] The memory 132 may also include a program / utility 1325 having a set (at least one) of program modules 1324, such program modules 1324 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0287] The electronic device 130 may also communicate with one or more external devices 134 (e.g., keyboards, pointing devices, etc.), may communicate with one or more devices that enable a user to interact with the electronic device 130, and / or communicate with any device that enables the electronic device 130 to communicate with one or more other electronic devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 135. Furthermore, the electronic device 130 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 136. As shown, the network adapter 136 communicates with other modules for the electronic device 130 via a bus 133. It should be understood that although Fig.14 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 130, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0288] In some possible implementations, various aspects of a heading angle determination method provided by the present application may also be implemented in the form of a program product, which includes a program code. When the program product is run on a computer device, the program code is used to enable the computer device to execute the steps of a heading angle determination method according to various exemplary implementations of the present application described above in this specification.

[0289] The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0290] The program product for heading angle determination of the embodiment of the present application can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on an electronic device. However, the program product of the present application is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0291] A readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, wherein readable program code is carried. Such propagated data signals may take a variety of forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination of the foregoing. A readable signal medium may also be any readable medium other than a readable storage medium, which may transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0292] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0293] The program code for performing the operations of the present application can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and also conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code can be executed entirely on the user electronic device, partially on the user device, as an independent software package, partially on the user electronic device and partially on the remote electronic device, or entirely on the remote electronic device or server. In the case of a remote electronic device, the remote electronic device can be connected to the user electronic device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external electronic device (for example, using an Internet service provider to connect through the Internet).

[0294] It should be noted that, although several units or subunits of the device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided into multiple units to be embodied.

[0295] In addition, although the operations of the method of the present application are described in a specific order in the drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0296] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0297] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0298] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0299] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0300] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for determining a heading angle, characterized in that: The method comprises: The target scene is continuously detected by a millimeter-wave radar to obtain a first detection point cluster of multiple consecutive frames, and the target scene is continuously detected by a laser radar to obtain a second detection point cloud of multiple consecutive frames; wherein the target scene includes at least one vehicle; Performing time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain a first detection point cluster and a second detection point cloud that are associated in both time and space; The coordinates of each first detection point in the first detection point cluster after synchronous processing are expanded according to the standard deviation to obtain an expanded area of ​​the first detection point, and the second detection point falling in the expanded area is determined according to the coordinates of the second detection point; wherein the standard deviation is the standard deviation of the physical parameter associated with the first detection point; wherein the coordinates of the first detection point include radial distance, azimuth and pitch angle, and the expanded area of ​​the radial distance, azimuth and pitch angle of each first detection point in the first detection point cluster after synchronous processing is determined according to the standard deviation; Associating a second detection point falling in the extended area with the first detection point, and fusing information of the mutually associated first detection point and second detection point to obtain a fused detection point; the fused detection point constitutes a fused detection point cloud; The fused detection point cloud composed of the fused detection points is segmented to obtain the fused detection point cloud corresponding to each vehicle in the target scene, and the following steps are performed for each frame of the fused detection point cloud of each vehicle: an initial heading angle of each vehicle is determined based on the fused detection point cloud; a target fused detection point is screened out according to the signal-to-noise ratio of each fused detection point in the fused detection point cloud; a wheel position difference of each vehicle is determined based on the fused detection point cloud and the fused detection point cloud of the previous frame; and the initial heading angle, the target fused detection point and the wheel position difference are input into a pre-trained motion model to obtain the heading angle of each vehicle.

2. The method according to claim 1, characterized in that The performing time synchronization processing on the first detection point cluster and the second detection point cloud includes: For each second timestamp, determine a first calculation timestamp and a second calculation timestamp corresponding to the second timestamp; wherein the first calculation timestamp is obtained by marking a first detection point cluster of each frame obtained by continuously detecting the target scene by the millimeter wave radar at a first preset frequency, and the second timestamp is obtained by marking a second detection point cloud of each frame obtained by continuously detecting the target scene by the laser radar at a second preset frequency; the first calculation timestamp is a first timestamp that is located before the second timestamp in time sequence and has the smallest time interval with the second timestamp, and the second calculation timestamp is a first timestamp that is located after the second timestamp in time sequence and has the smallest time interval with the second timestamp; An associated first detection point cluster having an association relationship with a second detection point cloud corresponding to the second timestamp is determined based on the first calculation timestamp, the second calculation timestamp, and the second timestamp.

3. The method according to claim 2, characterized in that The determining, based on the first calculation timestamp, the second calculation timestamp and the second timestamp, an associated first detection point cluster having an associated relationship with the second detection point cloud corresponding to the second timestamp comprises: Obtain a third timestamp, wherein the third timestamp is determined to be obtained by marking at a second preset frequency starting from a first time interval after the laser radar marks the first second timestamp; wherein the first time interval is obtained according to the first preset frequency; For each third timestamp, determine a target second timestamp that has the smallest timing difference with the third timestamp and is before the third timestamp; Obtain a first calculation timestamp and a second calculation timestamp corresponding to the target second timestamp; Determine an associated first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp; The associated first detection point cluster is used as a first detection point cluster that has a temporal association relationship with the second detection point cloud corresponding to the target second timestamp.

4. The method according to claim 3, characterized in that: The determining and associating the first detection point cluster based on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp includes: Performing linear interpolation processing on the first detection point cluster corresponding to the first calculation timestamp and the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster; or Mean processing is performed on the coordinates of the first detection point in the detection point cluster corresponding to the first calculation timestamp and the first detection point in the first detection point cluster corresponding to the second calculation timestamp to obtain the associated first detection point cluster.

5. The method according to claim 1, characterized in that: Performing spatial synchronization processing on the first detection point cluster and the second detection point cloud includes: The coordinates of the first detection point are transformed based on the rotation parameters and the translation parameters, and the coordinates of the second detection point are transformed based on the rotation parameters and the translation parameters to obtain the coordinates of the first detection point and the coordinates of the second detection point after spatial synchronization; wherein the rotation parameters are used to rotate the coordinate axes in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point to the left to coincide with the coordinate axes of the spatial synchronization coordinate system; and the translation parameters are used to determine the coordinate components of the origin of the spatial synchronization coordinate system in the coordinate system corresponding to the first detection point and the coordinate system corresponding to the second detection point.

6. The method according to claim 5, characterized in that The rotation parameters and the translation parameters are obtained according to the following method: Using a laser radar to measure a target object, obtain at least one measurement point of the target object and coordinates corresponding to the measurement point; and determine the mean value of the coordinates of the measurement point; Measuring the target object using a millimeter wave radar to obtain a first coordinate of the target object; A rotation parameter and a translation parameter are determined based on the mean of the coordinates and the first coordinate.

7. The method according to claim 1, characterized in that The step of determining the extended area of ​​the first detection point in radial distance, azimuth angle, and elevation angle of each first detection point in the first detection point cluster after synchronization processing according to the standard deviation comprises: The following process is performed for each first detection point in the first detection point cluster: Determine a radial distance standard deviation, an azimuth standard deviation, and an elevation standard deviation of the first detection point according to a signal-to-noise ratio of the first detection point; Determine a first difference between the radial distance of the first detection point and the standard deviation of the radial distance, and determine a first sum between the radial distance of the first detection point and the standard deviation of the radial distance; Determine an expansion area of ​​the radial distance according to the first difference and the first sum; Determine a second difference between the azimuth of the first detection point and the azimuth standard deviation, and determine a second sum between the azimuth of the first detection point and the azimuth standard deviation; Determine an extended area of ​​the azimuth according to the second difference and the second sum; Determine a third difference between the pitch angle of the first detection point and the standard deviation of the pitch angle, and determine a third sum between the pitch angle of the first detection point and the standard deviation of the pitch angle; Determining an extended area of ​​the pitch angle according to the third sum; An expansion area of ​​the first detection point is determined according to the expansion area of ​​the radial distance, the expansion area of ​​the azimuth angle, and the expansion area of ​​the elevation angle.

8. The method according to claim 1, characterized in that The associating the second detection point located in the extended area with the first detection point comprises: If a plurality of second detection points fall within the extended area of ​​the first detection point, determining a second detection point with the largest signal-to-noise ratio in the extended area; The second detection point with the largest signal-to-noise ratio is used as the second detection point associated with the first detection point.

9. The method according to claim 1, characterized in that: The first detection point cluster includes: a plurality of first detection points, the second detection point cloud includes: a plurality of second detection points, and the information fusion of the mutually related first detection points and second detection points to obtain fused detection points includes: For each second detection point, perform the following process: Using the physical parameter information of the first detection point associated with the second detection point as the physical parameter information of the fused detection point corresponding to the second detection point; Using the spatially synchronized coordinates of the second detection point as the coordinates of the fused detection point corresponding to the second detection point; The fused detection point is formed based on the physical parameter information of the fused detection point and the coordinates of the fused detection point.

10. A heading angle determination device, characterized in that: The device comprises: A detection module, configured to continuously detect the target scene using a millimeter wave radar to obtain a first detection point cluster of multiple consecutive frames, and to continuously detect the target scene using a laser radar to obtain a second detection point cloud of multiple consecutive frames; wherein the target scene includes at least one vehicle; A synchronization module, used for performing time synchronization processing and space synchronization processing on the first detection point cluster and the second detection point cloud to obtain a first detection point cluster and a second detection point cloud that are associated in both time and space; An expansion module, used for performing regional expansion on the coordinates of each first detection point in the first detection point cluster after synchronous processing according to a standard deviation to obtain an expanded area of ​​the first detection point, and determining a second detection point falling in the expanded area according to the coordinates of the second detection point; wherein the standard deviation is a standard deviation of a physical parameter associated with the first detection point; wherein the coordinates of the first detection point include a radial distance, an azimuth angle, and a pitch angle, and determining the expanded area of ​​the first detection point according to the standard deviation on the expanded area of ​​the radial distance, the azimuth angle, and the pitch angle of each first detection point in the first detection point cluster after synchronous processing; a fusion module, configured to associate a second detection point falling in the extended area with the first detection point, and to fuse information of the mutually associated first detection point and second detection point to obtain a fused detection point; the fused detection point constitutes a fused detection point cloud; The heading angle determination module is used to segment the fused detection point cloud composed of the fused detection points to obtain the fused detection point cloud corresponding to each vehicle in the target scene, and execute the following steps for each frame of the fused detection point cloud of each vehicle: determining the initial heading angle of each vehicle based on the fused detection point cloud; screening out the target fused detection point according to the signal-to-noise ratio of each fused detection point in the fused detection point cloud; determining the wheel position difference of each vehicle based on the fused detection point cloud and the fused detection point cloud of the previous frame; and inputting the initial heading angle, the target fused detection point and the wheel position difference into a pre-trained motion model to obtain the heading angle of each vehicle.

11. An electronic device, characterized in that: It comprises at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to implement the method as described in any one of claims 1 to 9.

12. A computer storage medium, characterized in that: The computer storage medium stores a computer program for enabling a computer to execute the method according to any one of claims 1 to 9.

Images