A Method and System for Calculating the Target Orientation Based on Millimeter-Wave Radar

Abstract

The present invention relates to the field of intelligent driving technology, and specifically discloses a method for calculating the target orientation based on a millimeter-wave radar, including obtaining the current frame point cloud information of the targets around the vehicle collected by the millimeter-wave radar; dividing the targets around the vehicle in the current frame into moving targets and stationary targets according to the current frame point cloud information of the targets around the vehicle and the vehicle's own speed information, and extracting the current frame point cloud information of the moving targets; performing clustering tracking on the current frame point cloud information of the moving targets to obtain the effective moving targets among all the moving targets in the current frame; extracting edge points of the effective moving targets in the current frame to obtain the edge point information of the effective moving targets in the current frame; and calculating the first orientation information of the effective moving targets in the current frame according to the edge point information of the effective moving targets in the current frame. The present invention also discloses a system for calculating the target orientation based on a millimeter-wave radar. The present invention can improve the accuracy of target orientation calculation.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent driving, and more specifically, to a method for calculating the target orientation based on a millimeter-wave radar and a system for calculating the target orientation based on a millimeter-wave radar. Background Art

[0002] With the continuous development of science and technology and the continuous pursuit of people for the living standard, there are more and more researches on the field of intelligent driving, and environmental perception is an important technology in intelligent driving.

[0003] Environmental perception mainly perceives the environment around the vehicle through various sensors, and obtains familiar information such as the position, speed orientation, size, shape, etc. of obstacles in the environment, so as to help with target recognition, clustering and tracking.

[0004] For the acquisition of the target orientation, most of the existing technologies judge through a deep learning training model based on a camera or a lidar. However, the recognition model based on a camera is greatly affected by weather, light, lane line clarity, and lane line quality, and the accuracy of the calculation based on the overhead image of the lidar is poor. Summary of the Invention

[0005] In order to solve the deficiencies existing in the prior art, the present invention provides a method and a system for calculating the target orientation based on a millimeter-wave radar, so as to solve the problems that the recognition model based on a camera in the related art is greatly affected by weather, light, lane line clarity, and lane line quality, and the accuracy of the calculation based on the overhead image of the lidar is poor.

[0006] As the first aspect of the present invention, a method for calculating the target orientation based on a millimeter-wave radar is provided, including:

[0007] Step S1: Obtain the current frame point cloud information of the targets around the vehicle collected by the millimeter-wave radar, where the current frame point cloud information of the targets around the vehicle includes the angle, radial velocity, and distance information of the targets around the vehicle in the current frame relative to the millimeter-wave radar;

[0008] Step S2: According to the current frame point cloud information of the targets around the vehicle and the vehicle's own speed information, divide the targets around the vehicle in the current frame into targets moving relative to the ground and targets stationary relative to the ground, and extract the current frame point cloud information of the targets moving relative to the ground from the current frame point cloud information of the targets around the vehicle;

[0009] Step S3: Cluster and track the current frame point cloud information of the targets moving relative to the ground to obtain valid moving targets among all the targets moving relative to the ground in the current frame;

[0010] Step S4: Extract edge points of the valid moving target in the current frame to obtain the edge point information of the valid moving target in the current frame;

[0011] Step S5: Calculate the first orientation information of the valid moving target in the current frame according to the edge point information of the valid moving target in the current frame.

[0012] Further, after the step S5, it further includes:

[0013] Obtain the motion speed information at both ends of the valid moving target in the current frame, and obtain the orientation information of the valid moving target in the previous frame;

[0014] Calculate the second orientation information of the valid moving target in the current frame and predict the orientation information of the valid moving target in the next frame according to the first orientation information of the valid moving target in the current frame and the motion speed information at both ends of the valid moving target in the current frame;

[0015] Calculate the true orientation information of the valid moving target in the current frame according to the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame.

[0016] Further, the calculating the true orientation information of the valid moving target in the current frame according to the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame further includes:

[0017] Compare the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame, and judge the true orientation information of the valid moving target in the current frame according to the comparison result among the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame;

[0018] If the difference among the three does not exceed the set threshold, the true orientation information of the valid moving target in the current frame is the second orientation information of the valid moving target in the current frame;

[0019] If the difference between the second orientation information of the valid moving target in the current frame and the orientation information of the valid moving target in the previous frame exceeds the set threshold, the true orientation information of the valid moving target in the current frame is the orientation information of the valid moving target in the previous frame.

[0020] Further, the step of clustering and tracking the current frame point cloud information of the ground moving targets to obtain the valid moving targets among all the ground moving targets in the current frame further includes:

[0021] Clustering the ground moving targets by using the density-based DBSCAN clustering algorithm, and extracting the trajectories of the ground moving targets by using dynamic Kalman filtering;

[0022] The clusters of ground moving target points that meet the density requirement threshold of the DBSCAN clustering algorithm and are tracked by continuous trajectories within a specified period of time are regarded as valid moving targets.

[0023] Further, the millimeter-wave radar is a 4D millimeter-wave radar.

[0024] As a second aspect of the present invention, a target orientation calculation system based on a millimeter-wave radar is provided, including:

[0025] A first acquisition module, configured to acquire the current frame point cloud information of the targets around the vehicle collected by the millimeter-wave radar, where the current frame point cloud information of the targets around the vehicle includes the angles, radial velocities, and distance information of the points of the targets around the vehicle in the current frame relative to the millimeter-wave radar;

[0026] A first extraction module, configured to divide the targets around the vehicle in the current frame into ground moving targets and ground stationary targets according to the current frame point cloud information of the targets around the vehicle and the vehicle's own speed information, and extract the current frame point cloud information of the ground moving targets from the current frame point cloud information of the targets around the vehicle;

[0027] A clustering and tracking module, configured to cluster and track the current frame point cloud information of the ground moving targets to obtain the valid moving targets among all the ground moving targets in the current frame;

[0028] A second extraction module, configured to extract the edge point information of the valid moving targets in the current frame by extracting the edge points of the valid moving targets in the current frame;

[0029] A first calculation module, configured to calculate the first orientation information of the valid moving targets in the current frame according to the edge point information of the valid moving targets in the current frame.

[0030] Further, it further includes a second acquisition module, a second calculation module, and a third calculation module;

[0031] The second acquisition module is configured to acquire the motion speed information at both ends of the valid moving targets in the current frame, and acquire the orientation information of the valid moving targets in the previous frame;

[0032] A second calculation module, configured to calculate a second orientation information of the valid moving target in the current frame and predict the orientation information of the valid moving target in the next frame according to the first orientation information of the valid moving target in the current frame and the moving speed information at both ends of the valid moving target in the current frame;

[0033] A third calculation module, configured to calculate the true orientation information of the valid moving target in the current frame according to the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame.

[0034] Further, the third calculation module is specifically configured to compare the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame, and judge the true orientation information of the valid moving target in the current frame according to the comparison result among the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame; if the difference among the three does not exceed a set threshold, the true orientation information of the valid moving target in the current frame is the second orientation information of the valid moving target in the current frame; if the difference between the second orientation information of the valid moving target in the current frame and the orientation information of the valid moving target in the previous frame exceeds the set threshold, the true orientation information of the valid moving target in the current frame is the orientation information of the valid moving target in the previous frame.

[0035] Further, the clustering and tracking module is specifically configured to cluster the ground moving targets through a density-based DBSCAN clustering algorithm and extract the trajectories of the ground moving targets through dynamic Kalman filtering;

[0036] A cluster of ground moving target points that meets the density requirement threshold of the DBSCAN clustering algorithm and is tracked by continuous trajectories within a specified period of time is regarded as a valid moving target.

[0037] Further, the millimeter-wave radar is a 4D millimeter-wave radar.

[0038] The method and system for calculating the target orientation based on the millimeter-wave radar provided by the present invention have the following advantages:

[0039] 1. Only use the millimeter-wave radar to calculate the drivable area of the road, without the cooperation of other sensors, not affected by weather, light, and ground, with good anti-interference ability, high stability, and strong robustness, and more applicable;

[0040] 2. Utilize the point cloud information output by the millimeter-wave radar and assist in judging the target orientation based on the velocity information of the target edge points extracted.

[0041] 3. Use the millimeter-wave radar to detect the trajectory of the target and assist in judging the orientation of the target in the front and rear frames, thereby improving the accuracy of the discriminant of the drivable area. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention, but do not constitute a limitation to the present invention.

[0043] Figure 1 It is a schematic diagram of the coordinate system in the present invention.

[0044] Figure 2 It is a flowchart of the method for calculating the target orientation based on the millimeter-wave radar in the present invention.

[0045] Figure 3 It is a flowchart of the specific implementation manner of the method for calculating the target orientation based on the millimeter-wave radar in the present invention. SPECIFIC EMBODIMENTS

[0046] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following describes in detail the specific implementation manners, structures, features, and effects of the method and system for calculating the target orientation based on the millimeter-wave radar according to the present invention in conjunction with the drawings and preferred embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0047] As Figure 1 shown, it is a schematic diagram of the coordinate system in the present invention. The direction parallel to the vehicle head is the X direction, with the left side being the negative direction and the right side being the positive direction. The direction perpendicular to the vehicle head is the Y direction, with the front being the positive direction and the rear being the negative direction. The millimeter-wave radar device involved in the present invention is installed at the front bumper position of the vehicle.

[0048] In this embodiment, a method for calculating the target orientation based on the millimeter-wave radar is provided. As Figure 2 shown, the method for calculating the target orientation based on the millimeter-wave radar includes:

[0049] Step S1: Obtain the current frame point cloud information of the target around the vehicle collected by the millimeter-wave radar, where the current frame point cloud information of the target around the vehicle includes the angle, radial velocity, and distance information of the target point around the vehicle in the current frame relative to the millimeter-wave radar;

[0050] It should be noted that the radar starts working. The radar's working mode is periodic. The radar system includes modules such as signal transmission, signal reception, and signal processing, and finally obtains the point cloud information of the target, including the angle, radial velocity, and distance of the target point relative to the millimeter wave radar. The duration of the above modules to complete one time is called a frame. It will be repeated continuously during the operation of the radar.

[0051] Step S2: dividing the vehicle surrounding objects in the current frame into ground moving objects and ground stationary objects according to the current frame point cloud information of the vehicle surrounding objects and the vehicle's own speed information, and extracting the current frame point cloud information of the ground moving objects from the current frame point cloud information of the vehicle surrounding objects;

[0052] Step S3: clustering and tracking the current frame point cloud information of the ground moving target to obtain the valid moving target among all the ground moving targets in the current frame;

[0053] In order to screen out all valid moving targets, the ground moving targets are clustered by the density-based DBSCAN clustering algorithm, and the trajectories of the ground moving targets are extracted by dynamic Kalman filtering; only the ground moving target point clusters that meet the density requirement threshold of the DBSCAN clustering algorithm and are tracked by continuous trajectories within a specified period of time will be regarded as valid moving targets.

[0054] Step S4: extracting edge points of the effective moving target in the current frame to obtain edge point information of the effective moving target in the current frame; wherein the edge point information of the effective moving target in the current frame includes longitudinal coordinates, transverse coordinates, pitch angles and motion speed information;

[0055] Step S5: Calculate the first orientation information of the effective moving target in the current frame according to the edge point information of the effective moving target in the current frame.

[0056] Specifically, the minimum circumscribed matrix of the polygon overlooking the edge point information of the target is calculated by the rotating caliper algorithm, and the first orientation of the target is determined.

[0057] The application basis of the rotating calipers algorithm here is that for a circumscribed rectangle of polygon P, there is a side that is collinear with the original polygon. At this time, using the rotating calipers algorithm, the edge point cloud polygon of the target is regarded as having two pairs of tangents tangent to the four endpoints in the x and y directions, and these four lines have determined a circumscribed rectangle of the polygon. Rotate this rectangle until it coincides with a side of the polygon, and then calculate and save the area of the current rectangle. Repeat the rotation traversal until the rotation angle is greater than 90°, and then output the coordinates of the rectangle with the smallest area, which is the first orientation of the target. It should be noted that there are two possible first orientations of the target at this time, so the following steps need to be carried out to further determine the true orientation of the target.

[0058] Preferably, as Figure 3 shown, after the step S5, it further includes:

[0059] Obtain the motion speed information at both ends of the effective moving target in the current frame, and obtain the orientation information of the effective moving target in the previous frame;

[0060] It should be noted that by comparing the speed difference between the farthest point and the nearest point of the effective moving target, combined with the coordinate information of the effective moving target, according to the working principle of the millimeter-wave radar, the speeds at both ends of the same target will vary according to the size of the target. After combining the speed information accumulated in a certain number of previous frames, auxiliary judgment can be made.

[0061] According to the first orientation information of the effective moving target in the current frame and the motion speed information at both ends of the effective moving target in the current frame, calculate the second orientation information of the effective moving target in the current frame and predict the orientation information of the effective moving target in the next frame;

[0062] According to the second orientation information of the effective moving target in the current frame, the orientation information of the effective moving target in the previous frame, and the orientation information of the effective moving target in the next frame, calculate the true orientation information of the effective moving target in the current frame.

[0063] Specifically, as Figure 3 shown, the calculating the true orientation information of the effective moving target in the current frame according to the second orientation information of the effective moving target in the current frame, the orientation information of the effective moving target in the previous frame, and the orientation information of the effective moving target in the next frame further includes:

[0064] Compare the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame, and judge the true orientation information of the valid moving target in the current frame according to the comparison results among the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame;

[0065] If the difference among the three does not exceed the set threshold, the true orientation information of the valid moving target in the current frame is the second orientation information of the valid moving target in the current frame;

[0066] If the difference between the second orientation information of the valid moving target in the current frame and the orientation information of the valid moving target in the previous frame exceeds the set threshold, the true orientation information of the valid moving target in the current frame is the orientation information of the valid moving target in the previous frame, and at the same time, save the second orientation information of the valid moving target in the current frame.

[0067] It should be noted that if the difference between the orientation information of the valid moving target in a certain number of subsequent frames and the second orientation information of the valid moving target in the current frame is less than the set threshold, the true orientation information of the valid moving target in the current frame is changed to the second orientation information of the valid moving target in the current frame.

[0068] Preferably, the millimeter-wave radar is a 4D millimeter-wave radar.

[0069] As another embodiment of the present invention, there is provided a target orientation calculation system based on a millimeter-wave radar, which includes:

[0070] A first acquisition module, configured to acquire the current frame point cloud information of the targets around the vehicle collected by the millimeter-wave radar, where the current frame point cloud information of the targets around the vehicle includes the angle, radial velocity, and distance information of the current frame of the targets around the vehicle relative to the millimeter-wave radar;

[0071] A first extraction module, configured to divide the current frame of the targets around the vehicle into targets moving relative to the ground and targets stationary relative to the ground according to the current frame point cloud information of the targets around the vehicle and the vehicle's own speed information, and extract the current frame point cloud information of the targets moving relative to the ground from the current frame point cloud information of the targets around the vehicle;

[0072] A clustering and tracking module, configured to perform clustering and tracking on the current frame point cloud information of the targets moving relative to the ground to obtain the valid moving targets among all the targets moving relative to the ground in the current frame;

[0073] A second extraction module, configured to extract edge points of the valid moving target in the current frame to obtain edge point information of the valid moving target in the current frame;

[0074] A first calculation module, configured to calculate first orientation information of the valid moving target in the current frame according to the edge point information of the valid moving target in the current frame.

[0075] Preferably, it further includes a second acquisition module, a second calculation module, and a third calculation module;

[0076] The second acquisition module is configured to acquire motion speed information at both ends of the valid moving target in the current frame, and acquire the orientation information of the valid moving target in the previous frame;

[0077] The second calculation module is configured to calculate second orientation information of the valid moving target in the current frame and predict the orientation information of the valid moving target in the next frame according to the first orientation information of the valid moving target in the current frame and the motion speed information at both ends of the valid moving target in the current frame;

[0078] The third calculation module is configured to calculate the true orientation information of the valid moving target in the current frame according to the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame.

[0079] Preferably, the third calculation module is specifically configured to compare the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame, and judge the true orientation information of the valid moving target in the current frame according to the comparison result among the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame; if the difference among the three does not exceed a set threshold, the true orientation information of the valid moving target in the current frame is the second orientation information of the valid moving target in the current frame; if the difference between the second orientation information of the valid moving target in the current frame and the orientation information of the valid moving target in the previous frame exceeds the set threshold, the true orientation information of the valid moving target in the current frame is the orientation information of the valid moving target in the previous frame.

[0080] Preferably, the clustering and tracking module is specifically configured to cluster the ground moving targets through the density-based DBSCAN clustering algorithm, and extract the trajectories of the ground moving targets through dynamic Kalman filtering;

[0081] A cluster of ground moving target points that meet the density requirement threshold of the DBSCAN clustering algorithm and are tracked by continuous trajectories within a specified period of time are regarded as valid moving targets.

[0082] Preferably, the millimeter-wave radar is a 4D millimeter-wave radar.

[0083] The present invention provides a method for calculating the target orientation based on a millimeter-wave radar. By using a single millimeter-wave radar sensor to calculate the target orientation, it can determine the target orientation by extracting target edge point information and judging the front and rear frames and speed. Using the millimeter-wave radar saves costs compared to lidar and is not affected by weather, lighting conditions, etc. like a camera. Moreover, by utilizing the characteristic that the radar detects the target to form a stable trajectory, the accuracy of target orientation calculation is improved.

[0084] The above are only the preferred embodiments of the present invention and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to obtain equivalent embodiments with equivalent changes. However, as long as the content does not depart from the technical solution of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A method for calculating the target orientation based on a millimeter-wave radar, characterized in that, Including: Step S1: Obtain the current frame point cloud information of the targets around the vehicle collected by the millimeter-wave radar, where the current frame point cloud information of the targets around the vehicle includes the angle, radial velocity, and distance information of the points of the targets around the vehicle in the current frame relative to the millimeter-wave radar; Step S2: According to the current frame point cloud information of the targets around the vehicle and the vehicle's own speed information, divide the targets around the vehicle in the current frame into ground-moving targets and ground-stationary targets, and extract the current frame point cloud information of the ground-moving targets from the current frame point cloud information of the targets around the vehicle; Step S3: Perform clustering tracking on the current frame point cloud information of the ground-moving targets to obtain the effective moving targets among all the ground-moving targets in the current frame; Step S4: Extract the edge points of the effective moving targets in the current frame to obtain the edge point information of the effective moving targets in the current frame; Step S5: Calculate the first orientation information of the effective moving targets in the current frame according to the edge point information of the effective moving targets in the current frame; Wherein, after the step S5, it further includes: Obtain the motion speed information at both ends of the effective moving targets in the current frame, and obtain the orientation information of the effective moving targets in the previous frame; According to the first orientation information of the effective moving targets in the current frame and the motion speed information at both ends of the effective moving targets in the current frame, calculate the second orientation information of the effective moving targets in the current frame and predict the orientation information of the effective moving targets in the next frame; According to the second orientation information of the effective moving targets in the current frame, the orientation information of the effective moving targets in the previous frame, and the orientation information of the effective moving targets in the next frame, calculate the true orientation information of the effective moving targets in the current frame.

2. The method for calculating the target orientation based on a millimeter-wave radar according to claim 1, wherein The calculating the true orientation information of the effective moving targets in the current frame according to the second orientation information of the effective moving targets in the current frame, the orientation information of the effective moving targets in the previous frame, and the orientation information of the effective moving targets in the next frame further includes: Compare the second orientation information of the effective moving targets in the current frame, the orientation information of the effective moving targets in the previous frame, and the orientation information of the effective moving targets in the next frame, and judge the true orientation information of the effective moving targets in the current frame according to the comparison result among the second orientation information of the effective moving targets in the current frame, the orientation information of the effective moving targets in the previous frame, and the orientation information of the effective moving targets in the next frame; If the difference among the three does not exceed the set threshold, the true orientation information of the effective moving targets in the current frame is the second orientation information of the effective moving targets in the current frame; If the difference between the second orientation information of the effective moving targets in the current frame and the orientation information of the effective moving targets in the previous frame exceeds the set threshold, the true orientation information of the effective moving targets in the current frame is the orientation information of the effective moving targets in the previous frame.

3. The method for calculating the target orientation based on a millimeter-wave radar according to claim 1, wherein The performing clustering tracking on the current frame point cloud information of the ground-moving targets to obtain the effective moving targets among all the ground-moving targets in the current frame further includes: Cluster the ground-moving targets through the density-based DBSCAN clustering algorithm, and extract the trajectories of the ground-moving targets through dynamic Kalman filtering; The clusters of ground-moving target points that meet the density requirement threshold of the DBSCAN clustering algorithm and are tracked by continuous trajectories within a specified period of time are regarded as valid moving targets.

4. The method for calculating the target orientation based on a millimeter-wave radar according to claim 1, characterized in that The millimeter-wave radar is a 4D millimeter-wave radar.

5. A target orientation calculation system based on a millimeter-wave radar, characterized in that, It includes: A first acquisition module, configured to acquire the current frame point cloud information of the targets around the vehicle collected by the millimeter-wave radar, where the current frame point cloud information of the targets around the vehicle includes the angle, radial velocity, and distance information of the targets around the vehicle in the current frame relative to the millimeter-wave radar; A first extraction module, configured to divide the targets around the vehicle in the current frame into ground-moving targets and ground-stationary targets according to the current frame point cloud information of the targets around the vehicle and the vehicle's own speed information, and extract the current frame point cloud information of the ground-moving targets from the current frame point cloud information of the targets around the vehicle; A clustering and tracking module, configured to perform clustering and tracking on the current frame point cloud information of the ground-moving targets to obtain the valid moving targets among all the ground-moving targets in the current frame; A second extraction module, configured to extract the edge point information of the valid moving targets in the current frame from the valid moving targets in the current frame; A first calculation module, configured to calculate the first orientation information of the valid moving targets in the current frame according to the edge point information of the valid moving targets in the current frame; Among them, it further includes a second acquisition module, a second calculation module, and a third calculation module; The second acquisition module is configured to acquire the motion speed information at both ends of the valid moving targets in the current frame, and acquire the orientation information of the valid moving targets in the previous frame; The second calculation module is configured to calculate the second orientation information of the valid moving targets in the current frame and predict the orientation information of the valid moving targets in the next frame according to the first orientation information of the valid moving targets in the current frame and the motion speed information at both ends of the valid moving targets in the current frame; The third calculation module is configured to calculate the true orientation information of the valid moving targets in the current frame according to the second orientation information of the valid moving targets in the current frame, the orientation information of the valid moving targets in the previous frame, and the orientation information of the valid moving targets in the next frame.

6. The target orientation calculation system based on millimeter wave radar according to claim 5, characterized in that, The third calculation module is specifically configured to compare the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame, and determine the true orientation information of the valid moving target in the current frame according to the comparison result among the second orientation information of the valid moving target in the current frame, the orientation information of the valid moving target in the previous frame, and the orientation information of the valid moving target in the next frame; if the difference among the three does not exceed the set threshold, the true orientation information of the valid moving target in the current frame is the second orientation information of the valid moving target in the current frame; if the difference between the second orientation information of the valid moving target in the current frame and the orientation information of the valid moving target in the previous frame exceeds the set threshold, the true orientation information of the valid moving target in the current frame is the orientation information of the valid moving target in the previous frame.

7. The target orientation calculation system based on a millimeter-wave radar according to claim 5, characterized in that, The clustering and tracking module is specifically configured to cluster the ground moving targets through the density-based DBSCAN clustering algorithm, and extract the trajectories of the ground moving targets through dynamic Kalman filtering; The ground moving target point clusters that meet the density requirement threshold of the DBSCAN clustering algorithm and have continuous trajectories tracked within a specified period of time are regarded as valid moving targets.

8. The target orientation calculation system based on a millimeter-wave radar according to claim 5, characterized in that, The millimeter-wave radar is a 4D millimeter-wave radar.

Images