Multi-path point cloud verification method and device, medium and equipment

Abstract

The invention provides a multipath point cloud verification method and device, a medium and equipment, and the method comprises the steps: carrying out the 360-degree surrounding sensing under the deployment of a plurality of vehicle-mounted millimeter wave radars, and enabling a plurality of radar detection overlapping regions to be included; when a target object is included in a certain radar detection overlapping area, obtaining detection results of two vehicle-mounted millimeter wave radars corresponding to the current radar detection overlapping area and predicted motion information of the target object; carrying out multipath point cloud verification based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter wave radars, and identifying the multipath point cloud; and when a target object is not included in a certain radar detection overlapping area, obtaining detection results of the two vehicle-mounted millimeter wave radars corresponding to the current radar detection overlapping area, performing preprocessing based on the detection results of the two vehicle-mounted millimeter wave radars, performing multipath point cloud verification based on the preprocessed detection results, and identifying the multipath point cloud.

Description

Technical Field

[0001] The present invention relates to the field of driverless technology, and more specifically, to a multi-path point cloud verification method, device, medium, and equipment. More specifically, it relates to a multi-path point cloud verification method, device, medium, and equipment based on the deployment of multiple vehicle-mounted millimeter-wave radars. Background Art

[0002] In practical applications, millimeter-wave radars may encounter ghosting phenomena, which are false targets caused by multi-path reflections or other interferences. Multi-path point cloud verification and ghosting suppression are one of the key technologies to improve the performance of radar systems.

[0003] After the millimeter-wave radar sensing technology has been gradually improved, 360-degree surround sensing can be achieved by deploying multiple millimeter-wave radars. However, there will still be missed detection phenomena in multi-path point cloud recognition through existing technologies, resulting in residual multi-path point clouds in the 360-degree surround point cloud output of multiple radars.

[0004] Therefore, the present invention provides a multi-path point cloud verification method and system based on the deployment of multiple vehicle-mounted millimeter-wave radars, which further identifies residual multi-path point clouds in 360-degree surround sensing under the deployment of multiple millimeter-wave radars. Summary of the Invention

[0005] The main purpose of the present invention is to solve the technical problem of missed detection in multi-path point cloud recognition in the prior art.

[0006] The first aspect of the present invention provides a multi-path point cloud verification method, including:

[0007] In 360-degree surround sensing under the deployment of multiple vehicle-mounted millimeter-wave radars, there are multiple radar detection overlapping areas;

[0008] When a target object is included in a certain radar detection overlapping area, obtain the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area and the predicted motion information of the target object; perform multi-path point cloud verification based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars to identify multi-path point clouds;

[0009] When a target object is not included in a certain radar detection overlapping area, obtain the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, perform preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars, and perform multi-path point cloud verification based on the preprocessed detection results to identify multi-path point clouds.

[0010] Optionally, in the first implementation manner of the first aspect of the present invention, when a target object is included in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area includes:

[0011] When a target object is included in a certain radar detection overlapping area, obtaining two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, including a first vehicle-mounted millimeter-wave radar and a second vehicle-mounted millimeter-wave radar;

[0012] Obtaining a corresponding point cloud result based on the first vehicle-mounted millimeter-wave radar, the point cloud result including: a first distance in the polar coordinate system and a first Doppler velocity vd1; and converting the first distance in the polar coordinate system from the radar polar coordinate system to the rear axle center point rectangular coordinate system to obtain the first distance in the rear axle center point rectangular coordinate system;

[0013] Obtaining a corresponding point cloud result based on the second vehicle-mounted millimeter-wave radar, the point cloud result including: a second distance in the polar coordinate system and a second Doppler velocity vd2, and converting the second distance in the polar coordinate system from the radar polar coordinate system to the rear axle center point rectangular coordinate system to obtain the second distance in the rear axle center point rectangular coordinate system;

[0014] Calculating the distance variance of the point cloud based on the first distance in the rear axle center point rectangular coordinate system and the second distance in the rear axle center point rectangular coordinate system.

[0015] Optionally, in the second implementation manner of the first aspect of the present invention, the predicted motion information of the target object includes: predicting the transverse and longitudinal velocities vx, vy and position information of the target object based on the fusion result of the multi-radar point cloud in the previous frame;

[0016] Judging whether the point clouds generated by the two radars are both within the preset range of the target object contour based on the predicted position information of the target object.

[0017] Optionally, in the third implementation manner of the first aspect of the present invention, performing multipath point cloud verification based on the predicted motion information of the target object and the detection results of two vehicle-mounted millimeter-wave radars to identify multipath point clouds, including:

[0018] Respectively calculating the radial velocities vr1, vr2 of two vehicle-mounted millimeter-wave radars corresponding to the radar detection overlapping area based on the predicted transverse and longitudinal velocities vx, vy of the target object;

[0019] When it is detected that the point clouds generated by two radars are both within the preset range of the target object contour and the distance variance of the point clouds generated by the two radars satisfies the preset range, calculate the variance E1 of the first Doppler velocity vd1 of the first radar and the radial velocity vr1 of the first radar; the variance E2 of the second Doppler velocity vd2 of the second radar and the radial velocity vr2 of the second radar;

[0020] When E1 < E2, E1 < the threshold, and E2 > the threshold, it indicates that the Doppler velocity of the point cloud of the second radar in the current overlapping area is an outlier, and the point cloud of the current second radar is identified as a multipath point cloud.

[0021] Optionally, in the fourth implementation manner of the first aspect of the present invention, when there is no target object in the detection overlapping area of a certain radar, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area includes:

[0022] When there is no target object in the detection overlapping area of a certain radar, obtain two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, including the first vehicle-mounted millimeter-wave radar and the second vehicle-mounted millimeter-wave radar;

[0023] Based on the first vehicle-mounted millimeter-wave radar, obtain the corresponding point cloud results, and the point cloud results include: the first distance and the first horizontal angle in the polar coordinate system;

[0024] Based on the second vehicle-mounted millimeter-wave radar, obtain the corresponding point cloud results, and the point cloud results include: the second distance and the second horizontal angle in the polar coordinate system.

[0025] Optionally, in the fifth implementation manner of the first aspect of the present invention, the preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars includes:

[0026] Based on the first distance and the first horizontal angle in the polar coordinate system obtained by the first vehicle-mounted millimeter-wave radar, convert them to the rear axle rectangular coordinate system of the vehicle, and obtain the first distance and the first horizontal angle in the rear axle rectangular coordinate system of the vehicle; then convert the first distance and the first horizontal angle in the rear axle rectangular coordinate system of the vehicle to the polar coordinate system of the second vehicle-mounted millimeter-wave radar, and obtain the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar.

[0027] Optionally, in the sixth implementation manner of the first aspect of the present invention, the multipath point cloud verification based on the preprocessed detection results to identify the multipath point cloud includes:

[0028] When the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar also exist within the field of view of the second vehicle-mounted millimeter-wave radar, and there is no target occlusion within the first distance and the first horizontal angle in the polar coordinate system of the current second vehicle-mounted millimeter-wave radar; the position where the current point cloud cluster is located can only be observed in the field of view of the first vehicle-mounted millimeter-wave radar and not in the field of view of the second vehicle-mounted millimeter-wave radar, then it is determined that the current point cloud cluster is multipath point cloud.

[0029] The second aspect of the present invention provides a multipath point cloud verification device, including:

[0030] A deployment module, used for 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, including multiple radar detection overlapping areas;

[0031] A multipath point cloud recognition module in the target object scenario, used for when a target object is included in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area and the predicted motion information of the target object; performing multipath point cloud verification based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars to identify multipath point cloud;

[0032] A multipath point cloud recognition module in the no-target object scenario, used for when no target object is included in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars, and performing multipath point cloud verification based on the preprocessed detection results to identify multipath point cloud.

[0033] Optionally, in the first implementation manner of the second aspect of the present invention, the obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area when a target object is included in a certain radar detection overlapping area includes:

[0034] When a target object is included in a certain radar detection overlapping area, obtaining two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, including a first vehicle-mounted millimeter-wave radar and a second vehicle-mounted millimeter-wave radar;

[0035] Based on the first vehicle-mounted millimeter-wave radar, obtaining the corresponding point cloud results, the point cloud results including: the first distance in the polar coordinate system and the first Doppler velocity vd1; and converting the first distance in the polar coordinate system from the radar polar coordinate system to the rear axle center point rectangular coordinate system to obtain the first distance in the rear axle center point rectangular coordinate system;

[0036] Based on the second vehicle-mounted millimeter-wave radar, the corresponding point cloud results are obtained. The point cloud results include: the second distance in the polar coordinate system and the second Doppler velocity vd2. The second distance in the polar coordinate system is converted from the radar polar coordinate system to the rectangular coordinate system of the rear axle center point, and the second distance in the rectangular coordinate system of the rear axle center point is obtained.

[0037] Based on the first distance in the rectangular coordinate system of the rear axle center point and the second distance in the rectangular coordinate system of the rear axle center point, the distance variance of the point cloud is calculated.

[0038] Optionally, in the second implementation manner of the second aspect of the present invention, the predicted motion information of the target object includes: predicting the lateral and longitudinal velocities vx, vy and position information of the target object based on the fusion result of the multi-radar point cloud of the previous frame.

[0039] Based on the predicted position information of the target object, it is determined whether the point clouds generated by the two radars are both within the preset range of the target object contour.

[0040] Optionally, in the third implementation manner of the second aspect of the present invention, the multi-path point cloud verification is performed based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars to identify the multi-path point cloud, including:

[0041] Based on the predicted lateral and longitudinal velocities vx, vy of the target object, the radial velocities vr1, vr2 of the two vehicle-mounted millimeter-wave radars corresponding to the radar detection overlapping area are calculated respectively.

[0042] When it is detected that the point clouds generated by the two radars are both within the preset range of the target object contour and the distance variance of the point clouds generated by the two radars satisfies the preset range, the variance E1 of the first Doppler velocity vd1 of the first radar and the radial velocity vr1 of the first radar is calculated; the variance E2 of the second Doppler velocity vd2 of the second radar and the radial velocity vr2 of the second radar is calculated.

[0043] When E1 < E2, E1 < the threshold, and E2 > the threshold, it indicates that the Doppler velocity of the point cloud of the second radar in the current overlapping area is an outlier, and the point cloud of the current second radar is identified as a multi-path point cloud.

[0044] Optionally, in the fourth implementation manner of the second aspect of the present invention, when a certain radar detection overlapping area does not include the target object, obtaining the detection results of the two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area includes:

[0045] When a certain radar detection overlapping area does not include the target object, the two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area are obtained, including the first vehicle-mounted millimeter-wave radar and the second vehicle-mounted millimeter-wave radar.

[0046] Based on the first vehicle-mounted millimeter-wave radar, the corresponding point cloud results are obtained. The point cloud results include: the first distance and the first horizontal angle in the polar coordinate system.

[0047] Based on the second vehicle-mounted millimeter-wave radar, the corresponding point cloud results are obtained. The point cloud results include: the second distance and the second horizontal angle in the polar coordinate system.

[0048] Optionally, in the fifth implementation manner of the second aspect of the present invention, the preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars includes:

[0049] The first distance and the first horizontal angle in the polar coordinate system obtained based on the first vehicle-mounted millimeter-wave radar are converted to the right-angle coordinate system of the rear axle of the vehicle, and the first distance and the first horizontal angle in the right-angle coordinate system of the rear axle of the vehicle are obtained; then the first distance and the first horizontal angle in the right-angle coordinate system of the rear axle of the vehicle are converted to the polar coordinate system of the second vehicle-mounted millimeter-wave radar, and the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar are obtained.

[0050] Optionally, in the sixth implementation manner of the second aspect of the present invention, the multi-path point cloud verification based on the preprocessed detection results to identify multi-path point clouds includes:

[0051] When the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar also exist within the field of view of the second vehicle-mounted millimeter-wave radar, and there is no target occlusion within the first distance and the first horizontal angle in the polar coordinate system of the current second vehicle-mounted millimeter-wave radar; the position where the current point cloud cluster is located can only be observed in the field of view of the first vehicle-mounted millimeter-wave radar and not observed in the field of view of the second vehicle-mounted millimeter-wave radar, then it is determined that the current point cloud cluster is a multi-path point cloud.

[0052] The third aspect of the present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned multi-path point cloud verification method are implemented.

[0053] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, the steps of the above-mentioned multi-path point cloud verification method are implemented.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. In the multi-radar architecture of the present invention, through the overlapping area of the two radar fields of view, the multi-path point clouds existing in a certain radar are identified and filtered, increasing the tracking stability of the fusion target;

[0056] 2. The performance of the vehicle-mounted millimeter-wave radar provided by the present invention in multi-path point cloud recognition has been significantly improved, enabling the system to achieve safer and more reliable vehicle autonomous driving and environmental perception. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Other features, objects, and advantages of the present invention will become more apparent by reading the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0058] Figure 1 FIG. 9 is a first flowchart of the multi-path point cloud verification method provided by the embodiment of the present invention.

[0059] Figure 2 FIG. 13 is a schematic diagram of 360-degree surround perception under the deployment of multiple millimeter-wave radars.

[0060] Figure 3 FIG. 17 is a schematic diagram of a target object included in a radar detection overlap area.

[0061] Figure 4 FIG. 21 is a schematic diagram of the Doppler velocity of a target object included in a radar detection overlap area.

[0062] Figure 5 FIG. 25 is a schematic diagram of no target object in a radar detection overlap area.

[0063] Figure 6 FIG. 29 is a schematic structural diagram of the multi-path point cloud verification device provided by the embodiment of the present invention.

[0064] Figure 7 FIG. 33 is a schematic structural diagram of the electronic device provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0065] The embodiment of the present invention provides a multi-path point cloud verification method, system, medium, and device. In the 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, it includes multiple radar detection overlap areas. When a target object is included in a certain radar detection overlap area, the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlap area and the predicted motion information of the target object are obtained. Based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars, multi-path point cloud verification is performed to identify multi-path point clouds. When no target object is included in a certain radar detection overlap area, the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlap area are obtained, preprocessing is performed based on the detection results of the current two vehicle-mounted millimeter-wave radars, and multi-path point cloud verification is performed based on the preprocessed detection results to identify multi-path point clouds. The present invention solves the technical problem of missed detection in multi-path point cloud recognition in the prior art.

[0066] In the description, claims and the above-mentioned drawings of the present invention, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the term "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0067] For ease of understanding, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , the first embodiment of the multipath point cloud verification method in the embodiments of the present invention includes:

[0068] 101. In the 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, there are multiple radar detection overlapping areas;

[0069] In this embodiment, in the 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, there are multiple radar detection overlapping areas. As Figure 2 shown, the horizontal angle field of view of the radar is in the range of 140-160 degrees. When six millimeter-wave radars are deployed, the overlapping area is as Figure 2 shown, and 360-degree surround perception can be achieved, where there is a large-angle overlapping area in the front area. And the front area is the key observation area for intelligent driving.

[0070] 102. When a target object is included in a certain radar detection overlapping area, obtain the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area and the predicted motion information of the target object; perform multipath point cloud verification based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars to identify multipath point clouds;

[0071] In this embodiment, when a target object is included in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area includes:

[0072] When a target object is included in a certain radar detection overlapping area, obtain two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, including a first vehicle-mounted millimeter-wave radar and a second vehicle-mounted millimeter-wave radar;

[0073] Obtain the corresponding point cloud results based on the first vehicle-mounted millimeter-wave radar. The point cloud results include: the first distance in the polar coordinate system and the first Doppler velocity vd1; and convert the first distance in the polar coordinate system from the radar polar coordinate system to the rectangular coordinate system of the rear axle center point to obtain the first distance in the rectangular coordinate system of the rear axle center point.

[0074] Obtain the corresponding point cloud results based on the second vehicle-mounted millimeter-wave radar. The point cloud results include: the second distance in the polar coordinate system and the second Doppler velocity vd2, and convert the second distance in the polar coordinate system from the radar polar coordinate system to the rectangular coordinate system of the rear axle center point to obtain the second distance in the rectangular coordinate system of the rear axle center point.

[0075] Calculate the distance variance of the point cloud based on the first distance in the rectangular coordinate system of the rear axle center point and the second distance in the rectangular coordinate system of the rear axle center point.

[0076] The predicted motion information of the target object includes: predicting the lateral and longitudinal velocities vx, vy and position information of the target object based on the fusion result of the multi-radar point cloud in the previous frame.

[0077] Based on the predicted position information of the target object, determine whether the point clouds generated by the two radars are both within the preset range of the target object contour.

[0078] The multi-path point cloud verification is performed based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars to identify the multi-path point cloud, including:

[0079] Based on the predicted lateral and longitudinal velocities vx, vy of the target object, calculate the radial velocities vr1, vr2 of the corresponding two vehicle-mounted millimeter-wave radars in the radar detection overlapping area respectively.

[0080] When it is detected that the point clouds generated by the two radars are both within the preset range of the target object contour and the distance variance of the point clouds generated by the two radars satisfies the preset range, calculate the variance E1 of the first Doppler velocity vd1 of the first radar and the radial velocity vr1 of the first radar; the variance E2 of the second Doppler velocity vd2 of the second radar and the radial velocity vr2 of the second radar.

[0081] When E1 < E2, E1 < threshold, and E2 > threshold, it means that the Doppler velocity of the point cloud of the second radar in the current overlapping area is an outlier, and identify the point cloud of the current second radar as a multi-path point cloud.

[0082] In this embodiment, as Figure 3 shown, when there is a target in the overlapping area of the two radars, the blue circle represents the point cloud detected by the front radar, the red circle represents the point cloud detected by the right front corner radar (multi-path point cloud), and the purple square represents the position after prediction of the fusion result of the multi-radar point cloud in the previous frame.

[0083] If the target is updated jointly by the front radar point cloud measurement value and the multipath point cloud of the right front corner radar at this time, a large amount of error will be introduced into the update of the target's motion state, causing the deviation of the target's motion state and even affecting the generation of the target in multiple future cycles.

[0084] However, at this time, since the target exists in the overlapping area of the two radar fields of view, the multipath point cloud of the right front corner radar can be identified by the detection results of the two radars plus the prediction result of the target.

[0085] The point cloud results detected by the radar include the distance, horizontal angle, and Doppler velocity (radial velocity) in polar coordinates. At this time, multiple radars are at different coordinate origins. After passing through the point cloud preprocessing module, the point cloud information is converted from the radar polar coordinate system to the right-angle coordinate system of the rear axle center point. At this time, the multi-radar point clouds are in the same coordinate system. However, since the detection method of the radar detects the radial velocity rather than the horizontal and vertical velocities in the rectangular coordinates, only the position of the point cloud can be converted to the rectangular coordinate system, and the velocity is still the Doppler radial velocity, as Figure 4 shown, along the dotted line direction, pointing to the corresponding radar.

[0086] Since the Doppler information origins of the point clouds detected by the two radars are not at the same place, the two Doppler vd1 and vd2 cannot be directly compared. At this time, the radial velocities vr1 and vr2 corresponding to the two radars can be deduced by fusing the horizontal and vertical velocities vx and vy of the target.

[0087] When it is detected that the point clouds generated by the two radars are both near the target contour and the distance variance of the two point clouds satisfies a certain range, at this time

[0088] Calculate the variance E1 between vd1 and vr1 of the front radar;

[0089] Calculate the variance E2 between vd2 and vr2 of the right front corner radar;

[0090] When E1 < E2, E1 < threshold and E2 > threshold are satisfied, it can be concluded that the Doppler velocity of the point cloud of the right front corner radar in this area is abnormal, that is, it is a multipath ghost. After that, these point clouds will not be used for subsequent clustering and tracking calculations.

[0091] 103. When the target object is not included in the detection overlapping area of a certain radar, obtain the detection results of the two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, perform preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars, and perform multipath point cloud verification based on the preprocessed detection results to identify the multipath point cloud.

[0092] In this embodiment, when a target object is not included in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area includes:

[0093] When a target object is not included in a certain radar detection overlapping area, obtaining two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, including a first vehicle-mounted millimeter-wave radar and a second vehicle-mounted millimeter-wave radar;

[0094] Based on the first vehicle-mounted millimeter-wave radar, obtaining a corresponding point cloud result, where the point cloud result includes: a first distance and a first horizontal angle in the polar coordinate system;

[0095] Based on the second vehicle-mounted millimeter-wave radar, obtaining a corresponding point cloud result, where the point cloud result includes: a second distance and a second horizontal angle in the polar coordinate system.

[0096] The preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars includes:

[0097] Based on the first distance and the first horizontal angle in the polar coordinate system obtained by the first vehicle-mounted millimeter-wave radar, converting them to the rear axle rectangular coordinate system of the vehicle, obtaining the first distance and the first horizontal angle in the rear axle rectangular coordinate system of the vehicle; then converting the first distance and the first horizontal angle in the rear axle rectangular coordinate system of the vehicle to the polar coordinate system of the second vehicle-mounted millimeter-wave radar, obtaining the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar.

[0098] The multipath point cloud verification based on the preprocessed detection results to identify multipath point clouds includes:

[0099] When the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar also exist within the field of view of the second vehicle-mounted millimeter-wave radar, and there is no target occlusion within the first distance and the first horizontal angle in the polar coordinate system of the current second vehicle-mounted millimeter-wave radar; the position where the current point cloud cluster is located can only be observed in the field of view of the first vehicle-mounted millimeter-wave radar and not observed in the field of view of the second vehicle-mounted millimeter-wave radar, then it is determined that the current point cloud cluster is a multipath point cloud.

[0100] In this embodiment, in the overlapping and unobstructed area, a single radar has multipath ghosts, such as Figure 5 shown, a multipath point cloud cluster that appears due to chaotic reflection of electromagnetic waves beside the guardrail.

[0101] The blue circles represent the point clouds (multipath point clouds) detected by the front radar. At this time, due to the lack of target prior information, this multipath point cloud cluster will form a new target start. If only a single radar (front radar) is used, it is impossible to suppress this multipath ghost.

[0102] However, since this point cloud exists in the overlapping area of the front radar and the left front corner radar and is not blocked in the fields of view of both radars, it can be judged through the following steps:

[0103] This point cloud is within the field of view of the front radar and there are no other targets in this horizontal angle.

[0104] According to the horizontal angle 1 and distance 1 detected by the front radar, the point cloud is converted from the polar coordinate system of the front radar to the rectangular coordinate system of the rear axle of the vehicle.

[0105] Then, this point cloud is converted to the left front corner coordinate system to obtain the corresponding distance 2 and horizontal angle 2.

[0106] It is judged through distance 2 and horizontal angle 2 that the point cloud also exists within the field of view of the left front corner and there is no target occlusion within this horizontal angle 2 and distance 2.

[0107] The position where this point cloud cluster is located is only observed in the field of view of the front radar and not observed in the left front corner radar (the left front corner radar only observes the static guardrail).

[0108] It is thus judged that this point cloud cluster belongs to multipath point cloud, and through the verification of two radars, a scenario that cannot be recognized by a single radar is achieved.

[0109] The multipath point cloud verification method in the embodiments of the present invention has been described above. Next, the multipath point cloud verification device in the embodiments of the present invention will be described. Please refer to Figure 6 One embodiment of the multipath point cloud verification device in the embodiments of the present invention includes:

[0110] The deployment module 201 is used for 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, including multiple radar detection overlapping areas;

[0111] In this embodiment, the deployment module 201 includes:

[0112] In 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, there are multiple radar detection overlapping areas. For example, Figure 2 As shown, the horizontal angle field of view of the radar is in the range of 140 - 160 degrees. When six millimeter-wave radars are deployed, the overlapping area is as shown in Figure 2 As shown, 360-degree surround perception can be achieved, and there is a large-angle overlapping area in the front area. And the front area is the key observation area for intelligent driving.

[0113] The multipath point cloud recognition module 202 in the target object scenario is used to obtain the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area and the predicted motion information of the target object when the target object is included in a certain radar detection overlapping area; perform multipath point cloud verification based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars, and identify the multipath point cloud.

[0114] In this embodiment, the multipath point cloud recognition module 202 in the target object scenario includes:

[0115] As Figure 3 shown, when there is a target in the overlapping area of two radars, the blue circle represents the point cloud detected by the front radar, the red circle represents the point cloud detected by the right front corner radar (multipath point cloud), and the purple square represents the predicted position of the fusion result of the multi-radar point cloud in the previous frame.

[0116] If the target is updated jointly by the point cloud measurement value of the front radar and the multipath point cloud of the right front corner radar at this time, a large amount of error will be introduced into the update of the target's motion state, causing the deviation of the target's motion state and even affecting the generation of the target in multiple future cycles.

[0117] However, at this time, since the target exists in the overlapping area of the fields of view of the two radars, the multipath point cloud of the right front corner radar can be identified by adding the detection results of the two radars and the prediction result of the target.

[0118] The point cloud results detected by the radar include the distance, horizontal angle, and Doppler velocity (radial velocity) in polar coordinates. At this time, multiple radars are at different coordinate origins. After passing through the point cloud preprocessing module, the point cloud information is converted from the radar polar coordinate system to the rear axle center point rectangular coordinate system. At this time, the multi-radar point clouds are in the same coordinate system. However, since the detected radial velocity by the radar detection method is not the horizontal and vertical velocities in the rectangular coordinates, only the position of the point cloud can be converted to the rectangular coordinate system, and the velocity is still the Doppler radial velocity, as Figure 4 shown, along the dotted line direction, pointing to the corresponding radar.

[0119] Since the Doppler information origins of the point clouds detected by the two radars are not at the same place, the two Doppler vd1 and vd2 cannot be directly compared. At this time, the radial velocities vr1 and vr2 corresponding to the two radars can be inversely deduced by fusing the horizontal and vertical velocities vx and vy of the target.

[0120] When it is detected that the point clouds generated by the two radars are both near the target contour and the distance variance between the two point clouds satisfies a certain range, at this time

[0121] Calculate the variance E1 between the front radar vd1 and vr1;

[0122] Calculate the variance E2 between the right front corner radar vd2 and vr2;

[0123] When E1 < E2, E1 < threshold and E2 > threshold are satisfied, it can be concluded that the Doppler velocity of the point cloud of the right front corner radar in this area is abnormal, that is, it is a multipath ghost. After that, these point clouds will not be used for subsequent clustering and tracking calculations.

[0124] The multipath point cloud recognition module 203 in the no-target object scenario is used to obtain the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area when the target object is not included in a certain radar detection overlapping area, preprocess based on the detection results of the current two vehicle-mounted millimeter-wave radars, and perform multipath point cloud verification based on the preprocessed detection results to identify multipath point clouds.

[0125] In this embodiment, the multipath point cloud recognition module 203 in the no-target object scenario includes:

[0126] In the overlapping and unobstructed area, multipath ghosts appear in a single radar, such as Figure 5 As shown, a cluster of multipath point clouds appears due to chaotic reflection of electromagnetic waves beside the guardrail.

[0127] The blue circle represents the point cloud (multipath point cloud) detected by the front radar. At this time, due to the lack of prior target information, this multipath point cloud group will form a new target start. If only through a single radar (front radar), it is impossible to suppress this multipath ghost.

[0128] However, since this point cloud exists in the common overlapping area of the front radar and the left front corner radar and there is no occlusion in the fields of view of both radars, it can be judged through the following steps:

[0129] This point cloud is within the field of view of the front radar and there are no other targets in this horizontal angle

[0130] According to the horizontal angle 1 and distance 1 detected by the front radar, convert the point cloud from the polar coordinate system of the front radar to the rear axle rectangular coordinate system of the vehicle

[0131] Then convert this point cloud to the left front corner coordinate system to obtain the corresponding distance 2 and horizontal angle 2

[0132] Judge through distance 2 and horizontal angle 2 that the point cloud also exists within the field of view of the left front corner and there is no target occlusion within this horizontal angle 2 and distance 2

[0133] The position where this point cloud cluster is located is only observed in the field of view of the front radar and not observed in the left front corner radar (the left front corner radar only observes the static guardrail)

[0134] It can be judged from this that this point cloud cluster belongs to multipath point clouds, and through the verification of two radars, a scenario that cannot be recognized by a single radar is achieved.

[0135] above Figure 6 The multi-path point cloud verification device in the embodiments of the present invention is described in detail from the perspective of modular functional entities. Next, the electronic device in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0136] Figure 7 FIG. is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. The electronic device 700 may vary greatly due to different configurations or performances, and may include one or more central processing units (CPUs) 710 (for example, one or more processors) and a memory 720, and one or more storage media 730 (for example, one or more mass storage devices) for storing application programs 733 or data 732. Among them, the memory 720 and the storage media 730 may be transient storage or persistent storage. The program stored in the storage media 730 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the electronic device 700. Further, the processor 710 may be configured to communicate with the storage media 730 and execute a series of instruction operations in the storage media 730 on the electronic device 700.

[0137] The electronic device 700 may further include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input / output interfaces 750, and / or one or more operating systems 731, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on. Those skilled in the art can understand that Figure 7 the shown structure of the electronic device does not constitute a limitation on the electronic device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0138] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are run on a computer, the computer is caused to execute the steps of the multi-path point cloud verification method.

[0139] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices, or units can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0140] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0141] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of various embodiments of the present invention.

Claims

1. A multipath point cloud verification method, characterized in that, Including: In the 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, there are multiple radar detection overlapping areas; When a target object is included in a certain radar detection overlapping area, obtain the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area and the predicted motion information of the target object; Perform multipath point cloud verification based on the predicted motion information of the target object and the detection results of two vehicle-mounted millimeter-wave radars to identify multipath point clouds; When a target object is not included in a certain radar detection overlapping area, obtain the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, perform preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars, and perform multipath point cloud verification based on the preprocessed detection results to identify multipath point clouds.

2. The multipath point cloud verification method according to claim 1, characterized in that The step of, when a target object is included in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area includes: When a target object is included in a certain radar detection overlapping area, obtain two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, including a first vehicle-mounted millimeter-wave radar and a second vehicle-mounted millimeter-wave radar; Obtain the corresponding point cloud result based on the first vehicle-mounted millimeter-wave radar. The point cloud result includes: the first distance in the polar coordinate system and the first Doppler velocity vd1; and convert the first distance in the polar coordinate system of the radar to the rectangular coordinate system of the rear axle center point to obtain the first distance in the rectangular coordinate system of the rear axle center point; Obtain the corresponding point cloud result based on the second vehicle-mounted millimeter-wave radar. The point cloud result includes: the second distance in the polar coordinate system and the second Doppler velocity vd2, and convert the second distance in the polar coordinate system of the radar to the rectangular coordinate system of the rear axle center point to obtain the second distance in the rectangular coordinate system of the rear axle center point; Calculate the distance variance of the point cloud based on the first distance in the rectangular coordinate system of the rear axle center point and the second distance in the rectangular coordinate system of the rear axle center point.

3. The multi-path point cloud verification method according to claim 2, wherein The predicted motion information of the target object includes: predicting the lateral and longitudinal velocities vx, vy and position information of the target object based on the fusion result of the multi-radar point cloud in the previous frame; Judge whether the point clouds generated by the two radars are both within the preset range of the target object contour based on the predicted position information of the target object.

4. The multi-path point cloud verification method according to claim 3, wherein The step of performing multipath point cloud verification based on the predicted motion information of the target object and the detection results of two vehicle-mounted millimeter-wave radars to identify multipath point clouds includes: Calculate the radial velocities vr1, vr2 of two vehicle-mounted millimeter-wave radars corresponding to the radar detection overlapping area respectively based on the predicted lateral and longitudinal velocities vx, vy of the target object; When it is detected that the point clouds generated by the two radars are both within the preset range of the target object contour and the distance variance of the point clouds generated by the two radars meets the preset range, calculate the variance E1 of the first Doppler velocity vd1 of the first radar and the radial velocity vr1 of the first radar; the variance E2 of the second Doppler velocity vd2 of the second radar and the radial velocity vr2 of the second radar; When E1 < E2, E1 < the threshold, and E2 > the threshold, it indicates that the Doppler velocity of the point cloud of the second radar in the current overlapping area is an outlier, and the point cloud of the current second radar is identified as multipath point cloud.

5. The multipath point cloud verification method according to claim 1, wherein When there is no target object in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area includes: When there is no target object in a certain radar detection overlapping area, obtaining two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, including the first vehicle-mounted millimeter-wave radar and the second vehicle-mounted millimeter-wave radar; Based on the first vehicle-mounted millimeter-wave radar, obtaining the corresponding point cloud results, where the point cloud results include: the first distance and the first horizontal angle in the polar coordinate system; Based on the second vehicle-mounted millimeter-wave radar, obtaining the corresponding point cloud results, where the point cloud results include: the second distance and the second horizontal angle in the polar coordinate system.

6. The multipath point cloud verification method according to claim 5, characterized in that, The preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars includes: Converting the first distance and the first horizontal angle in the polar coordinate system obtained based on the first vehicle-mounted millimeter-wave radar to the rear axle rectangular coordinate system of the vehicle, obtaining the first distance and the first horizontal angle in the rear axle rectangular coordinate system of the vehicle; then converting the first distance and the first horizontal angle in the rear axle rectangular coordinate system of the vehicle to the polar coordinate system of the second vehicle-mounted millimeter-wave radar, obtaining the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar.

7. The multipath point cloud verification method according to claim 6, characterized in that The multipath point cloud verification based on the preprocessed detection results to identify multipath point cloud includes: When the first distance and the first horizontal angle in the polar coordinate system of the second vehicle-mounted millimeter-wave radar also exist within the field of view of the second vehicle-mounted millimeter-wave radar, and there is no target occlusion within the first distance and the first horizontal angle in the polar coordinate system of the current second vehicle-mounted millimeter-wave radar; the current point cloud cluster can only be observed in the field of view of the first vehicle-mounted millimeter-wave radar and not in the field of view of the second vehicle-mounted millimeter-wave radar, then it is determined that the current point cloud cluster is a multipath point cloud.

8. A multipath point cloud verification device, characterized in that, Includes: A deployment module, used for 360-degree surround perception under the deployment of multiple vehicle-mounted millimeter-wave radars, including multiple radar detection overlapping areas; A multipath point cloud identification module in the target object scenario, used for when there is a target object in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area and the predicted motion information of the target object; Performing multipath point cloud verification based on the predicted motion information of the target object and the detection results of the two vehicle-mounted millimeter-wave radars to identify multipath point cloud; A multipath point cloud identification module in the no-target object scenario, used for when there is no target object in a certain radar detection overlapping area, obtaining the detection results of two vehicle-mounted millimeter-wave radars corresponding to the current radar detection overlapping area, performing preprocessing based on the detection results of the current two vehicle-mounted millimeter-wave radars, and performing multipath point cloud verification based on the preprocessed detection results to identify multipath point cloud.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the multipath point cloud verification method described in any one of claims 1 to 7.

10. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the computer program is executed by a processor, it implements the steps of the multipath point cloud verification method described in any one of claims 1 to 7.

Images