Laser radar signal processing device

Abstract

The present invention provides a laser radar signal processing device that can accurately estimate the horizontal axis misalignment angle of a laser radar. [Solution] The laser radar signal processing device 3 includes a point cloud information acquisition unit 31 that acquires information on a point cloud PG, which is a group of detection points included in a predetermined row from a plurality of detection points arranged in a matrix obtained by the laser radar scanning a predetermined detection range in two dimensions in directions perpendicular and horizontal to the road surface; a relative velocity acquisition unit 33 that acquires the relative velocity of the point cloud PG; a reference point extraction unit 34 that extracts a reference point from the point cloud PG based on the relative velocity of the point cloud PG; and an angle estimation unit 35 that estimates the axial misalignment angle of the laser radar in the horizontal direction based on the orientation of the reference point.

Description

[Technical Field]

[0001] This disclosure relates to a laser radar signal processing device. [Background technology]

[0002] When using laser radar mounted on a vehicle for object detection, misalignment of the laser radar's mounting angle may reduce object detection performance. For example, misalignment of the mounting angle may cause the radar to mistakenly identify a distant vehicle traveling in an adjacent lane as a vehicle in the same lane on a highway. It may also cause inconsistencies with the output of other sensors such as cameras and millimeter-wave radar. Laser radar is also known as LiDAR. LiDAR is an abbreviation for Light Detection and Ranging or Laser Imaging Detection and Ranging.

[0003] Therefore, it is necessary to install laser radars with precision at the vehicle factory. For example, when detecting an object in front of the vehicle, the laser radar must be installed so that the central axis of the detection range faces the front of the vehicle. However, even if the laser radar is installed with precision at the vehicle factory, the installation angle may shift due to aging of the equipment or external impacts after the vehicle is shipped, so a technology to detect the installation angle of the laser radar after the vehicle has been shipped is desirable. In practice, sensors such as laser radars are sometimes intentionally installed in directions other than facing the front of the vehicle, but in any case, there is a desired installation angle.

[0004] For example, Patent Document 1 creates a road surface plane based on reflected light from three points on the road surface and estimates the vertical axis misalignment of the laser radar by comparing the road surface plane with a reference plane. Furthermore, for horizontal axis misalignment, it detects white lines with high reflectivity on the road surface and estimates the horizontal axis misalignment angle based on the detection results of the white lines. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2022-150929 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] However, because the white lines have width, methods that estimate horizontal axis displacement based on the detection results of the white lines may result in large estimation errors.

[0007] In view of the above, this disclosure aims to provide a laser radar signal processing device that can estimate the horizontal axis misalignment angle of a laser radar with high accuracy. [Means for solving the problem]

[0008] To achieve the above objective, according to one aspect of this disclosure, the laser radar signal processing device (3) includes a point cloud information acquisition unit (31) that acquires information on a point cloud PG, which is a group of detection points included in a predetermined row from a plurality of detection points arranged in a matrix obtained by the laser radar (2) two-dimensionally scanning a predetermined detection range in a direction perpendicular to the road surface and in a direction horizontal to the road surface; a relative velocity acquisition unit (33) that acquires the relative velocity of the point cloud PG; a reference point extraction unit (34) that extracts a reference point from the point cloud PG based on the relative velocity of the point cloud PG; and an angle estimation unit (35) that estimates the axial misalignment angle of the laser radar in the horizontal direction based on the orientation of the reference point.

[0009] When a laser radar is mounted on a vehicle or other moving object, the relative velocity of a stationary object such as a road surface to the laser radar varies depending on the angle from the direction of movement of the laser radar. Therefore, by setting a point cloud PG to correspond to the road surface, the angle of the laser radar in the direction horizontal to the road surface can be estimated based on the relative velocity of the detection points included in the point cloud PG. Since this relative velocity changes significantly depending on the orientation of the detection point, the horizontal axis misalignment angle of the laser radar can be estimated with higher accuracy compared to estimating the laser radar angle based on the detection results of a wide white line.

[0010] Note that reference numerals in parentheses attached to each component or the like indicate an example of correspondence between the component or the like and specific components or the like described in the embodiments to be described later. Brief Description of Drawings

[0011] [Figure 1] It is a diagram showing a configuration of an in-vehicle system according to a first embodiment. [Figure 2] It is a diagram showing a detection range of a laser radar. [Figure 3] It is a diagram showing a two-dimensional scanning direction of emitted light of laser radars arranged in a matrix. [Figure 4] It is a diagram for explaining the relationship between the azimuth and relative velocity of a stationary target. [Figure 5] It is a diagram showing a detection range when there is no axis misalignment of the laser radar. [Figure 6] It is a diagram showing relative velocities of detection points when there is no axis misalignment of the laser radar. [Figure 7] It is a diagram showing a detection range when there is axis misalignment of the laser radar. [Figure 8] It is a diagram showing relative velocities of detection points when there is axis misalignment of the laser radar. [Figure 9] It is a block diagram of a laser radar signal processing device. [Figure 10] It is a flowchart of axis misalignment angle estimation processing. Modes for Carrying Out the Invention

[0012] Hereinafter, embodiments of the present disclosure will be described based on the drawings. Note that in the following respective embodiments, parts that are identical or equivalent to each other will be described with the same reference numerals assigned thereto.

[0013] (First Embodiment) The first embodiment will now be described. The in-vehicle system 1 shown in Figure 1 is mounted on a vehicle such as an automobile and detects objects around the vehicle. Hereinafter, the vehicle equipped with the in-vehicle system 1 will be referred to as "the vehicle." The object detection results from the in-vehicle system 1 are transmitted to an ECU (not shown) and used for vehicle control such as warnings to the vehicle's occupants and driving assistance. ECU stands for Electronic Control Unit.

[0014] As shown in Figure 1, the in-vehicle system 1 includes a laser radar 2 and a laser radar signal processing device 3. The laser radar 2 detects objects by scanning a predetermined detection range around the vehicle in two dimensions. This detection range is set to include the road surface. Specifically, the laser radar 2 is mounted near the top of the windshield, for example, in front of the rearview mirror, so as to face the front of the vehicle.

[0015] Laser radar 2 comprises a laser light source and an optical waveguide connected to the light source. The optical waveguide branches into multiple paths, and a phase shifter is provided at each branch to form a phased array antenna. The laser light generated by the light source is emitted to the outside as emitted light from the phased array antenna. Laser radar 2 scans the emitted light horizontally by phase control using the phase shifter and scans the emitted light vertically by wavelength control of the laser light. These horizontal and vertical directions are horizontal and perpendicular to the road surface. As mentioned above, laser radar 2 is mounted facing the front of the vehicle, and the emitted light is scanned horizontally in the left-right direction of the vehicle by phase control and vertically in the up-down direction by wavelength control.

[0016] As a result, a two-dimensional scan is performed in the detection range R in front of the vehicle 100, as shown in Figure 2. The emitted light is emitted at multiple angles in both the horizontal and vertical directions, thereby acquiring information from multiple detection points. For example, the emitted light is emitted at several hundred different angles in both the horizontal and vertical directions. In Figure 2, the white circles inside the detection range R indicate detection points. A detection point is a point on an object that is presumed to have reflected the emitted light at each angle.

[0017] Figure 2 illustrates the case where the emitted light illuminates the road surface across the entire detection range R, and detection points corresponding to the road surface are acquired. In this case, the vertical scanning of the emitted light corresponds to scanning in the depth direction in front of the vehicle 100. The detection range R is a fan-shaped area centered on the vehicle 100, with multiple detection points arranged on multiple arcs extending horizontally.

[0018] The laser radar 2 generates detection point cloud data based on the scan results and detects objects based on the detection point cloud data. As shown in Figure 3, the detection point cloud data is an image-like structure in which multiple detection points are arranged in a matrix in two dimensions corresponding to the horizontal and vertical directions, and it includes positional information in these two dimensions, as well as measurement results of distance and velocity.

[0019] Among the detection point clouds arranged in a matrix, the detection point clouds included in a predetermined row are defined as point cloud PG. As will be described later, the data of point cloud PG is used to estimate the axial misalignment angle of the laser radar 2, and it is desirable to set point cloud PG to consist of detection points corresponding to the road surface. The portion of the detection range R corresponding to the road surface changes depending on the installation environment of the laser radar 2, but in most cases, the vertical mounting angle of the LiDAR sensor is set so that the bottom row of the detection point cloud is closest to the road surface. Therefore, in this embodiment, the bottom row of the detection point cloud is defined as point cloud PG. The point cloud PG of the bottom row is the detection point cloud whose vertical position is closest to the vehicle 100 in the detection range R shown in Figure 2, that is, the detection point cloud located closest to the vehicle 100 as seen from the vehicle 100. The detection points at the left end and right end of point cloud PG are defined as P L , P RThe detection point located at the center of the point cloud PG is defined as P C Let's assume that.

[0020] Laser radar 2 employs the FMCW method, measuring the distance between laser radar 2 and each detection point, as well as the relative velocity of each detection point with respect to laser radar 2 (hereinafter simply referred to as relative velocity). FMCW stands for Frequency Modulated Continuous Wave. The detection point cloud data includes the aforementioned two-dimensional position information, distance information, and relative velocity information. Laser radar 2 is connected to laser radar signal processing device 3 via an in-vehicle communication line and transmits the detection point cloud data to the laser radar signal processing device 3.

[0021] The laser radar signal processing unit 3 performs target recognition processing for objects around the vehicle based on information acquired from the laser radar 2. The laser radar signal processing unit 3 is connected to an ECU (not shown) via an in-vehicle communication line and transmits the results of the target recognition processing to the ECU. The processing results from the laser radar signal processing unit 3 are used for vehicle control operations such as warning operations to the occupants and collision avoidance operations.

[0022] The laser radar signal processing unit 3 has a microcomputer configuration that includes a CPU, RAM, ROM, and non-volatile rewritable memory (not shown). CPU stands for Central Processing Unit. RAM stands for Random access memory. ROM stands for Read Only Memory. The laser radar signal processing unit 3 reads and executes a computer program stored in the ROM or non-volatile rewritable memory, which are non-transitional physical recording media. When this computer program is executed, the method corresponding to the computer program is performed. That is, the laser radar signal processing unit 3 performs various control processes, such as the control process shown in Figure 10, according to its computer program.

[0023] The relationship between the position of a stationary target and the relative velocity will be described. As shown in Figure 4, let the speed of the host vehicle 100, that is, the speed of the laser radar 2, be V EGO , and let the relative velocity of a stationary target B at an angle θ from the front of the host vehicle be V TGT , then V TGT =V EGO cosθ. Therefore, the relative velocity V TGT reaches its maximum value when the stationary target B is at an angle of θ=0°, that is, when it is in front of the host vehicle 100, and at this time V TGT =V EGO . The relative velocity V TGT decreases as the angle θ moves away from 0°. Note that the vertical component of the velocity is ignored here.

[0024] As shown in Figure 5, it is assumed that there is no axis misalignment of the laser radar 2, and there is no horizontal angular difference between the central axis of the detection range R and the traveling direction of the host vehicle 100. In this case, if the point group PG corresponds to the road surface, from the above relationship between V TGT and the angle θ, the relative velocity of the point group PG is as shown in Figure 6, and the relative velocity of the detection point P C is maximized.

[0025] On the other hand, as shown in Figure 7, when there is a horizontal axis misalignment of the laser radar 2, and there is a horizontal angular difference between the central axis of the detection range R indicated by the straight line L and the traveling direction of the host vehicle 100, the relative velocity of the point group PG is as shown in Figure 8, and the relative velocity of a detection point different from the detection point P C is maximized. Let the detection point at which the relative velocity is maximized be P MAX , and let the azimuth of the detection point P MAX , specifically, the angle from the central axis of the detection range R, be α. The angle α corresponds to the horizontal axis misalignment angle of the laser radar 2. That is, when the detection point P MAX is located at a position of angle α with one side in the left-right direction from the front of the laser radar 2 defined as positive, the horizontal axis misalignment angle of the laser radar 2 is -α with the one side defined as positive from the front of the host vehicle.

[0026] In this embodiment, this is used to estimate the angle of the laser radar 2. Specifically, as shown in Figure 9, the laser radar signal processing device 3 includes a point cloud information acquisition unit 31, a road surface determination unit 32, a relative velocity acquisition unit 33, a reference point extraction unit 34, and an angle estimation unit 35.

[0027] The point cloud information acquisition unit 31 acquires the detected point cloud data generated by the laser radar 2. Specifically, the point cloud information acquisition unit 31 acquires information on the distance from the laser radar 2 and the relative velocity for each detected point in the point cloud PG, which is the point closest to the road surface in the vertical direction, from the bottom row of the detected point cloud arranged in a matrix in the vertical and horizontal directions.

[0028] The road surface determination unit 32 determines, based on the detected point cloud data acquired by the point cloud information acquisition unit 31, whether the detected point cloud corresponds to the road surface and not to a preceding vehicle or the like. Specifically, the road surface determination unit 32 determines whether all the detected points included in the point cloud PG are points on the road surface. For example, the road surface determination unit 32 determines that the point cloud PG corresponds to the road surface if the distance from the laser radar 2 to all the detected points in the point cloud PG is within a predetermined reference value range. This determination method is effective when the mounting height of the laser radar 2 is predetermined.

[0029] Furthermore, if the emitted light is directed at an object other than the road surface, such as a preceding vehicle or bicycle, the distance between the laser radar 2 and the detection point corresponding to that object will deviate significantly from the distance between the laser radar 2 and other detection points in the point cloud PG. This can be used to determine that the point cloud PG corresponds to the road surface by applying outlier testing to the distance between the laser radar 2 and each detection point in the point cloud PG, and if there are no outliers. The distance between the laser radar 2 and the point cloud PG corresponding to the road surface is determined by the mounting height of the laser radar 2 and the emission angle of the emitted light, and the mounting height of the laser radar 2 varies depending on the vehicle type. In contrast, the determination method using outlier testing can determine whether or not it is the road surface even if the mounting height of the laser radar 2 changes.

[0030] The relative velocity acquisition unit 33 acquires the relative velocity of the point cloud PG from the detected point cloud data acquired by the point cloud information acquisition unit 31. Specifically, the relative velocity acquisition unit 33 calculates the average value of the relative velocities acquired N times for each detected point of the point cloud PG, after the point cloud information acquisition unit 31 has acquired the detected point cloud data N times, with N being a natural number. Although N=1 is acceptable, it is preferable to set N≧2 to reduce the influence of noise-induced errors in relative velocity.

[0031] The reference point extraction unit 34 extracts reference points from the point cloud PG based on the relative velocity of the point cloud PG. These reference points are detection points in the point cloud PG that are considered to have the maximum relative velocity. Specifically, the reference point extraction unit 34 extracts detection points P, which are the detection points at both ends of the point cloud PG in the horizontal direction. L , P R The point with the greater relative velocity is designated as the first point P1. Next, the reference point extraction unit 34 divides the point cloud PG into two point clouds, left and right, in the center. The detection point cloud containing the first point P1 is designated as the first point cloud PG1, and the other detection point cloud is designated as the second point cloud PG2. The detection point in the second point cloud PG2 whose relative velocity is closest to that of the first point P1 is designated as the second point P2. Next, the reference point extraction unit 34 selects the detection point in the point cloud PG that is closest to the midpoint between the first point P1 and the second point P2 as the reference point P S Extract it as follows.

[0032] Furthermore, if the number of detection points included in the point cloud PG is odd, then detection point P C It may belong to both the first point group PG1 and the second point group PG2, or to only one of them. Also, if there are two detection points in the point group PG that are closest to the midpoint between the first point P1 and the second point P2, one of the detection points will be used as the reference point P S Alternatively, a reference point P could be placed midway between these two detection points. S You may set it to that.

[0033] Furthermore, instead of extracting the first point P1, etc., the relative velocities of all detected points included in the point cloud PG are compared, and the detected point with the maximum relative velocity is selected as the reference point P. SIt may be extracted as follows. However, near the vertex of the cosine curve, the difference in relative velocity between two adjacent detection points is small. Therefore, detection point P with a large difference in relative velocity L , P R In the method of selecting the first point P1 and estimating the axial misalignment angle, the influence of noise is reduced and the accuracy of estimating the axial misalignment angle is improved compared to the method of comparing the relative velocities of all detected points in the point cloud PG.

[0034] Also, reference point P S The detection point is P MAX This corresponds to the detection point P. MAX Since the relative velocity does not necessarily coincide with the peak of the cosine curve, the detection point closest to the peak of the cosine curve is the reference point P. S This is how it is extracted. Quantization errors occur during this process, but the impact of these errors can be reduced, for example, by acquiring several hundred detection points in the horizontal direction.

[0035] The angle estimation unit 35 uses the reference point P extracted by the reference point extraction unit 34. S Based on the orientation of the reference point P, the angle of the laser radar 2 is estimated. Specifically, the angle estimation unit 35 estimates the angle of the reference point P. S The angle α is obtained, and angle -α is estimated as the horizontal axis misalignment angle of the laser radar 2. Based on the angle -α estimated by the angle estimation unit 35, the laser radar signal processing unit 3 corrects information such as the object detection result obtained from the laser radar 2, uses the corrected information to perform object recognition processing around the vehicle, and transmits the result of the object recognition processing to the ECU, etc.

[0036] The operation of this embodiment will now be described. In the in-vehicle system 1, a two-dimensional scan of the detection range R is performed periodically by the laser radar 2. The laser radar signal processing device 3 then periodically performs the following: estimation of the axial misalignment angle of the laser radar 2, and target recognition using the estimated axial misalignment angle.

[0037] The axial misalignment angle estimation process performed by the laser radar signal processing device 3 will be explained using Figure 10. First, in step S101, the laser radar signal processing device 3 acquires the latest detection point cloud data generated by the laser radar 2.

[0038] In the following step S102, the laser radar signal processing device 3 uses the method described above to determine whether the point cloud PG acquired in step S101 corresponds to the road surface. If it is determined that the point cloud PG corresponds to the road surface, the process proceeds to step S103. If it is determined that the point cloud PG does not correspond to the road surface, the process proceeds to step S101.

[0039] In step S103, the laser radar signal processing device 3 stores the detected point cloud data acquired in step S101. In the following step S104, the laser radar signal processing device 3 determines whether the number of times the detected point cloud data has been stored in step S103 is N or greater. If it is determined that the number is N or greater, the process proceeds to step S105. If it is determined that the number is not N or greater, the process proceeds to step S101.

[0040] In step S105, the laser radar signal processing device 3 calculates the average value of the relative velocities stored N times in step S103 for each detection point included in the point cloud PG. In the following step S106, the laser radar signal processing device 3 uses the method described above to select a reference point P from the point cloud PG based on the average value of the relative velocities calculated in step S105. S Extracts the reference point P. In the following step S107, the laser radar signal processing device 3 extracts the reference point P. S The angle α is obtained, and the angle -α is estimated as the axial misalignment angle. After step S107, the process ends.

[0041] In step S105, if a relative velocity calculation result like the one shown in Figure 8 is obtained, then detection point P L The relative velocity is detected at point P R Because it is greater than the relative velocity, detection point P in step S106 LThis becomes the first point P1. The left and right halves of the point cloud PG become the first point cloud PG1 and the second point cloud PG2, respectively, and the detection point in the second point cloud PG2 that has the closest relative velocity to the first point P1 becomes the second point P2. Furthermore, the detection point closest to the midpoint between the first point P1 and the second point P2 is the reference point P S Therefore, in step S107, the reference point P S The angle α is obtained, and the angle -α is estimated as the axial misalignment angle of laser radar 2.

[0042] As described above, in this embodiment, the angle of the laser radar 2 is estimated based on the relative velocity of the detection points included in the point cloud PG. Since the relative velocity changes significantly depending on the orientation of the detection points, the horizontal axis misalignment angle of the laser radar 2 can be estimated with high accuracy.

[0043] Furthermore, according to the above embodiment, the following effects can be obtained.

[0044] (1) The detection range of the laser radar 2 is set to include the road surface, and the point cloud PG is the detected point cloud of the row closest to the road surface. Since the emitted light is more likely to hit the road surface in the lower part of the detection range, setting the bottom row of the detected point cloud as the point cloud PG makes it easier for the point cloud PG to correspond to the road surface. Therefore, the frequency of the determination result of step S102 being YES can be increased, and the processing time for estimating the axis misalignment angle can be shortened.

[0045] (2) Whether the point cloud PG corresponds to the road surface is determined based on whether the distance between the laser radar 2 and each detection point included in the point cloud PG is within a predetermined reference range. This allows for the determination of whether or not it is the road surface in a simple manner.

[0046] (3) An outlier test is applied to the distance between the laser radar 2 and each detection point included in the point cloud PG to determine whether the point cloud PG corresponds to the road surface. This makes it possible to determine whether it is the road surface even if the mounting height of the laser radar 2 changes.

[0047] (4) Detection point P L , P RThe point with the greater relative velocity is designated as the first point P1. When the point cloud PG is divided into two point clouds in the middle, the point cloud containing the first point P1 is designated as the first point cloud PG1, and the other point cloud is designated as the second point cloud PG2. The detection point in the second point cloud PG2 whose relative velocity is closest to the first point P1 is designated as the second point P2, and the detection point in the point cloud PG that is closest to the midpoint between the first point P1 and the second point P2 is designated as the reference point P S This is how it is extracted. Compared to the method of comparing the relative velocities of all detected points in the point cloud PG, this reduces the influence of noise and improves the accuracy of estimating the axial misalignment angle.

[0048] (5) A reference point P is established based on the average value of the relative velocity obtained multiple times for each detection point in the point cloud PG. S This extracts the relevant information. This reduces the impact of noise-induced errors in relative velocity and improves the accuracy of axial misalignment angle estimation.

[0049] (Other embodiments) Furthermore, this disclosure is not limited to the embodiments described above and can be modified as appropriate. Also, it goes without saying that, in the embodiments described above, the elements constituting the embodiments are not necessarily essential unless explicitly stated as particularly essential or considered fundamentally essential. Furthermore, in the embodiments described above, when numerical values ​​such as the number, numerical values, quantities, or ranges of the components of the embodiments are mentioned, the embodiments are not limited to those specific numbers unless explicitly stated as particularly essential or considered fundamentally limited to a specific number. Also, when the shapes, positional relationships, etc., of the components are mentioned in the embodiments described above, the embodiments are not limited to those shapes, positional relationships, etc., unless explicitly stated or considered fundamentally limited to specific shapes, positional relationships, etc.

[0050] Laser radar 2 may employ the TOF method. TOF stands for Time of Flight. However, considering the accuracy of relative velocity measurement, it is desirable for laser radar 2 to employ the FMCW method.

[0051] In the first embodiment, if it was determined in step S102 that the point cloud PG did not correspond to the road surface, the process proceeded to step S101. That is, if the point cloud PG contains detection points that do not correspond to the road surface, the point cloud PG is not used to estimate the axial misalignment angle of the laser radar 2. However, detection points that do not correspond to the road surface may be removed from the point cloud PG, and the axial misalignment angle may be estimated using the remaining detection points.

[0052] Alternatively, a predetermined number of detection points may be removed from both ends of the point cloud PG, and the axial misalignment angle may be estimated using the remaining detection points. In this case, the leftmost and rightmost detection points of the remaining point cloud will be detection point P, respectively. L , P R This suppresses the inclusion of detection points of objects outside the lane, such as oncoming vehicles or curbs, in the point cloud PG, increasing the frequency of the judgment result in step S102 being YES and shortening the processing time for estimating the axis misalignment angle.

[0053] The detected point cloud data may be defined as a point cloud PG (Point Cloud Programming) from a predetermined number of rows away from the bottom row of the detected point cloud data. For example, any row within 20% of the bottom row of the detected point cloud data may be defined as the point cloud PG. This suppresses the inclusion of the hood in the point cloud PG and allows for accurate estimation of the axial misalignment angle.

[0054] Depending on the processing capacity of the laser radar signal processing device 3, the detection point cloud PG may be a set of detection points included in a predetermined number of rows of the matrix-like detection point cloud. This improves the accuracy of the relative velocity calculation.

[0055] In the first embodiment, the laser radar signal processing device 3 is separate from the laser radar 2. However, the laser radar signal processing device 3 may be located inside the laser radar 2. [Explanation of Symbols]

[0056] 2. Laser radar 3. Laser radar signal processing device 31 Point cloud information acquisition unit 33 Relative velocity acquisition unit 34 Reference Point Extraction Section 35° Angle Estimation Section

Claims

1. Laser radar signal processing device (3), A laser radar (2) performs a two-dimensional scan of a predetermined detection range in both perpendicular and horizontal directions on the road surface, and a point cloud information acquisition unit (31) acquires information on the point cloud PG, which is a group of detection points included in a predetermined row from among the multiple detection points arranged in a matrix obtained by the laser radar (2). A relative velocity acquisition unit (33) that acquires the relative velocity of the point cloud PG, A reference point extraction unit (34) extracts reference points from the point cloud PG based on the relative velocity of the point cloud PG, A laser radar signal processing device comprising: an angle estimation unit (35) that estimates the axial misalignment angle of the laser radar in the horizontal direction based on the orientation of the reference point.

2. The detection range is set to include the road surface. The laser radar signal processing device according to claim 1, wherein the predetermined row is the row closest to the road surface.

3. The detection range is set to include the road surface. The laser radar signal processing device according to claim 1, wherein the predetermined row is a predetermined number of rows located from the row closest to the road surface.

4. The laser radar signal processing device according to claim 1, wherein the point cloud PG is a group of detection points included in a predetermined number of rows from a plurality of detection points arranged in a matrix.

5. The system includes a road surface determination unit (32) that determines whether the point cloud PG corresponds to the road surface, The laser radar signal processing device according to claim 1, wherein the road surface determination unit determines whether the distance between the laser radar and the point cloud PG is within a predetermined reference value range.

6. The system includes a road surface determination unit (32) that determines whether the point cloud PG corresponds to the road surface, The laser radar signal processing device according to claim 1, wherein the road surface determination unit performs the determination by applying outlier testing to the distance to the point cloud PG.

7. The laser radar signal processing device according to claim 1, wherein the reference point extraction unit extracts the detection point with the largest relative velocity to the laser radar from the point cloud PG as a reference point.

8. Of the detection points at both ends of the point cloud PG, the one with the greater relative velocity is designated as the first point. When the point cloud PG is divided into two detection point clouds at the center, the detection point cloud containing the first point is designated as the first point cloud, and the other detection point cloud is designated as the second point cloud. Among the second group of points, the detection point whose relative velocity is closest to the first point is designated as the second point. The laser radar signal processing device according to claim 1, wherein the reference point extraction unit extracts the detection point from the point cloud PG that is closest to the midpoint between the first point and the second point as a reference point.

9. The relative velocity acquisition unit calculates the average value of the relative velocity acquired multiple times for each detection point of the point cloud PG, The laser radar signal processing device according to claim 1, wherein the reference point extraction unit extracts reference points based on the average value.

10. The laser radar signal processing device according to claim 1, which is separate from the aforementioned laser radar.

11. The laser radar signal processing device according to claim 1, which is provided inside the laser radar.

Images