LiDAR calibration method

Abstract

This invention discloses a lidar calibration method applied to lidar devices with reflective elements positioned within their scanning area. The calibration method includes dividing point cloud data into reflective virtual image point clouds and non-reflective point clouds based on the relative positions of the reflective elements and the lidar device. The reflective virtual image point clouds are collected by scanning laser light reflected from the reflective elements. Based on the principles of optical reflection and the initial reflective surface equation of the reflective elements, coordinate transformation is performed on each virtual image sampling point in the reflective virtual image point cloud to obtain real image sampling points. The nearest neighbor sampling point to the real image sampling point is searched in the non-reflective point cloud, and a target function is constructed based on the deviation between the real image sampling point and its corresponding neighbor sampling point. The parameters in the initial reflective surface equation are optimized iteratively to obtain calibration parameters that make the target function less than a target threshold. This application ensures high-precision scanning measurement of the lidar device, achieving efficient utilization of the lidar equipment.

Description

Technical Field

[0001] This invention relates to the field of lidar scanning technology, and in particular to a lidar calibration method. Background Technology

[0002] A lidar device is a physical performance testing instrument that achieves high-precision three-dimensional measurement by emitting laser pulses and receiving reflected signals. Its core components include a laser emitter, a receiver, a drive scanning mechanism, and a data processing system. The drive scanning mechanism drives the laser emitter and receiver to rotate 360°, thus achieving a large-area 360° scan.

[0003] However, in real-world applications, such a large scanning range of 3D measurement data may not be necessary. For example, LiDAR devices used in automobiles may only perform 3D measurements of obstacles or vehicles directly in front of them. In other words, during the scanning and acquisition of point cloud data, depending on the application scenario, some scanning data may not be used but still require a certain amount of computing power.

[0004] Therefore, avoiding the computational waste of data obtained by LiDAR scanning areas of no interest is of great significance for the efficient utilization of LiDAR equipment. Summary of the Invention

[0005] The purpose of this invention is to provide a lidar calibration method that ensures high-precision scanning measurement of lidar equipment while achieving efficient utilization of the equipment.

[0006] To address the aforementioned technical problems, this invention provides a laser radar calibration method, wherein a reflective element is provided within the scanning area of ​​the laser radar device to reflect laser light in a portion of the area; the calibration method includes:

[0007] Based on the relative position between the reflective element and the lidar device, the point cloud data collected by the lidar device is divided into reflected virtual image point cloud and non-reflective point cloud; wherein, the reflected virtual image point cloud is collected by scanning the laser light reflected by the reflective element;

[0008] Based on the principle of optical reflection and the initial reflective surface equation of the reflective element, coordinate transformation is performed on each virtual image sampling point in the reflected virtual image point cloud to obtain real image sampling points;

[0009] Search for the nearest neighboring sampling point to the real image sampling point in the non-reflection point cloud, and construct an objective function based on the deviation between the real image sampling point and the corresponding neighboring sampling point;

[0010] The parameters in the initial reflective surface equation are optimized and iterated to obtain calibration parameters that make the objective function less than the target threshold.

[0011] In an optional embodiment of this application, based on the relative position between the reflective element and the lidar device, the point cloud data collected by the lidar device is divided into reflective virtual image point clouds and non-reflective point clouds, including:

[0012] Based on the relative position between the reflective element and the lidar device, the scanning area of ​​the lidar device is divided into a reflective scanning angle area and a non-reflective scanning angle area;

[0013] The point cloud data acquired through the reflection scanning angle region is divided into the reflection virtual image point cloud, and the point cloud data acquired through the non-reflection scanning angle region is divided into the non-reflection point cloud.

[0014] In one optional embodiment of this application, the reflective element includes a first planar reflector and a second planar reflector; wherein one end of the first planar reflector and the second planar reflector are spliced ​​together, and the angle between the reflective surfaces is within a set angle range;

[0015] Based on the relative position between the reflective element and the lidar device, the scanning area of ​​the lidar device is divided into a reflective scanning angle region and a non-reflective scanning angle region, including:

[0016] Based on the relative position between the reflective element and the lidar device, the middle scanning direction, the first boundary direction, and the second boundary direction of the lidar device for scanning the splicing position between the first planar reflector and the second planar reflector and the mutually opposite side positions are determined;

[0017] The scanning angle region between the first boundary direction and the intermediate scanning direction, closer to the lidar device, is taken as the initial first reflection angle region.

[0018] The scanning angle region between the intermediate scanning direction and the second boundary direction is used as the initial second reflection angle region;

[0019] The initial first reflection angle region and the initial second reflection angle region are both reduced from their two end boundaries by a fine-tuning angle to determine the first reflection angle region and the second reflection angle region.

[0020] The first reflection angle region and the second reflection angle region are used as the reflection scanning angle region, and the scanning angle region on the side away from the lidar device between the first boundary direction and the second boundary direction is used as the non-reflection scanning angle region.

[0021] In an optional embodiment of this application, when the first planar reflector and the second planar reflector are symmetrically arranged about the lidar device, and the angle between the reflective surfaces is less than 180 degrees, the first reflection angle region and the second reflection angle region are used as the reflection scanning angle region, including:

[0022] Based on the relative position between the first planar reflector and the lidar, the first reflection angle region is divided into a first primary reflection angle region and a first secondary reflection angle region;

[0023] Based on the relative position between the second planar reflector and the lidar, the second reflection angle region is divided into a second primary reflection angle region and a second secondary reflection angle region.

[0024] The first primary reflection angle region, the first secondary reflection angle region, the second primary reflection angle region, and the second secondary reflection angle region are collectively referred to as the reflection scanning angle region;

[0025] Accordingly, the point cloud data acquired through the reflection scanning angle region is divided into the reflection virtual image point cloud, including:

[0026] The point cloud data output through the first primary reflection angle region is divided into a first primary reflection virtual image point cloud, the point cloud data output through the first secondary reflection angle region is divided into a first secondary reflection virtual image point cloud, the point cloud data output through the second primary reflection angle region is divided into a second primary reflection virtual image point cloud, and the point cloud data output through the second secondary reflection angle region is divided into a second secondary reflection virtual image point cloud.

[0027] In an optional embodiment of this application, according to the principle of optical reflection and the initial reflective surface equation of the reflective element, coordinate transformation is performed on each virtual image sampling point in the reflected virtual image point cloud to obtain real image sampling points, including:

[0028] Based on the relative positional relationship between the lidar device and the first planar reflector, and according to the reflective surface equation model... The initial equation of the first reflective surface of the first planar reflector in a three-dimensional Cartesian coordinate system is constructed; wherein the origin of the three-dimensional Cartesian coordinate system coincides with the center of the lidar device; wherein, All are coefficient parameters;

[0029] Based on the relative positional relationship between the lidar device and the second planar reflector, and according to the reflective surface equation model... Construct the initial equation of the second reflective surface of the second planar reflector in the three-dimensional rectangular coordinate system;

[0030] Based on the initial first reflective surface equation, the initial second reflective surface equation, and the reflection conversion matrix model, a first reflection conversion matrix and a second reflection conversion matrix are constructed; wherein, the reflection conversion matrix model is determined in advance based on the principle of optical reflection;

[0031] The first primary reflection virtual image point cloud is transformed using the first reflection transformation matrix. The first secondary reflection virtual image point cloud is transformed using the first reflection transformation matrix and the second reflection transformation matrix in sequence. The second primary reflection virtual image point cloud is transformed using the second reflection transformation matrix in sequence. The second secondary reflection virtual image point cloud is transformed using the second reflection transformation matrix and the first reflection transformation matrix in sequence to obtain real image sampling points.

[0032] In an optional embodiment of this application, the predetermined reflection transformation matrix model is: ;in, , , And the unit normal vector .

[0033] In an optional embodiment of this application, the reflective element includes a first plane mirror and a second plane mirror spliced ​​together with a reflective surface angle of less than 180 degrees, and a third plane mirror and a fourth plane mirror respectively disposed on both sides of the first plane mirror and the second plane mirror; the first plane mirror and the third plane mirror, as well as the second plane mirror and the fourth plane mirror, are spliced ​​together with a reflective surface angle of greater than 180 degrees and less than 270 degrees.

[0034] Accordingly, based on the relative position between the reflective element and the lidar device, the scanning area of ​​the lidar device is divided into a reflective scanning angle region and a non-reflective scanning angle region, including:

[0035] Based on the relative positions of the first, second, third, and fourth planar reflectors and the lidar, the non-reflective scanning angle region, as well as the first, second, third, and fourth reflection angle regions within the reflective scanning angle region, are determined.

[0036] In an optional embodiment of this application, the reflective surface angle between the first plane mirror and the second plane mirror, the reflective surface angle between the first plane mirror and the third plane mirror, and the reflective surface angle between the third plane mirror and the fourth plane mirror are all adjustable by a drive motor.

[0037] Accordingly, the process of determining the relative position between the reflective element and the lidar device, before separating the reflective virtual image point cloud and the non-reflective point cloud from the point cloud data collected by the lidar device, further includes:

[0038] Based on the driving parameters of the drive motor for the first, second, third, and fourth planar reflectors, the relative positions of the first, second, third, and fourth planar reflectors and the lidar device are determined.

[0039] In an optional embodiment of this application, searching for the nearest neighboring sampling point in the non-reflective point cloud that is closest to the real image sampling point includes:

[0040] For each real image sampling point, a sampling point with the smallest Euclidean distance to the real image sampling point in the non-reflection point cloud and whose Euclidean distance is not greater than a set distance threshold is searched and selected as the neighboring sampling point corresponding to the real image sampling point.

[0041] In an optional embodiment of this application, constructing an objective function for the deviation between the real image sampling point and the corresponding neighboring sampling point includes:

[0042] The objective function is constructed based on the distance deviation between the real image sampling point and the corresponding neighboring sampling point. ;in, This represents the parameters to be optimized in the initial reflective surface equation. This represents the total number of real image sampling points. Indicates in the parameter Under the conditions One real image sampling point; Indicates the first The neighboring sampling points corresponding to each real image sampling point.

[0043] This invention provides a lidar calibration method. The lidar device has a reflective element within its scanning area that reflects laser light in a portion of the area. The calibration method includes: dividing the point cloud data collected by the lidar device into a reflected virtual image point cloud and a non-reflective point cloud based on the relative position between the reflective element and the lidar device; wherein the reflected virtual image point cloud is collected by scanning laser light reflected by the reflective element; performing coordinate transformation on each virtual image sampling point in the reflected virtual image point cloud according to the optical reflection principle and the initial reflective surface equation of the reflective element to obtain real image sampling points; searching for the nearest neighbor sampling point to the real image sampling point in the non-reflective point cloud; constructing an objective function based on the deviation between the real image sampling point and its corresponding neighbor sampling point; and optimizing and iterating the parameters in the initial reflective surface equation to obtain calibration parameters that make the objective function less than a target threshold.

[0044] In this application, a reflective element is provided in the scanning area of ​​the lidar device. By using the reflective element to reflect and deflect the laser light from the non-interest area, the laser light from the lidar device can be more concentrated and radiated towards the interest area for scanning. This increases the laser scanning density in the interest area, thereby improving scanning accuracy and enhancing the efficient utilization of the lidar device.

[0045] Based on this, the laser light reflected by the reflective element needs to be based on the precise relative positional relationship between the reflective element and the lidar device to achieve accurate scanning measurement. Therefore, this application divides the point cloud data collected by the lidar device into reflected virtual image point clouds and non-reflective point clouds. Based on the principle of optical reflection, the reflected virtual image point clouds are transformed to obtain real image sampling points. Finally, based on the deviation between the real image sampling points and the non-reflective point clouds in the scanning results of the same scanning area, the initial reflective surface equation characterizing the relative position between the reflective element and the lidar device is optimized and iterated. Ultimately, the parameters characterizing the relative position between the reflective element and the lidar device are accurately calibrated, providing possible data basis for the lidar device to combine with the reflective element to accurately scan and measure the region of interest. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 A schematic diagram of a scanning area provided by the lidar device in the embodiments of this application;

[0048] Figure 2 for Figure 1 A schematic diagram of the LiDAR device within the scanning area of ​​the XOZ plane;

[0049] Figure 3 A schematic diagram of the scanning optical path for one arrangement of the reflective elements in a lidar device provided in an embodiment of this application;

[0050] Figure 4 A schematic flowchart illustrating the lidar calibration method provided in this application embodiment;

[0051] Figure 5 A schematic diagram illustrating the relative positional relationship between virtual image sampling points and real image sampling points provided in an embodiment of this application;

[0052] Figure 6 A schematic diagram illustrating one method of dividing the reflection angle region according to an embodiment of this application;

[0053] Figure 7 A schematic diagram illustrating another method of dividing the reflection angle region provided in an embodiment of this application;

[0054] Figure 8 A schematic diagram of the scanning optical path for another layout of the reflective element in the lidar device provided in the embodiments of this application;

[0055] Figure 9 A schematic diagram of the scanning optical path for another layout of the reflective element in the lidar device provided in the embodiments of this application;

[0056] In the attached diagram: 10 is a lidar device, 20 is a reflective element, 21 is a first plane reflector, 22 is a second plane reflector, 23 is a third plane reflector, and 24 is a fourth plane reflector. Detailed Implementation

[0057] The core of this invention is to provide a lidar calibration method that combines lidar equipment and reflective elements. This allows the lidar equipment's scanning light to perform targeted and concentrated scanning of the region of interest, and effectively calibrates the parameters characterizing the relative position between the lidar equipment and the reflective elements. This provides reliable data for achieving accurate scanning measurements with the lidar equipment, thereby improving scanning measurement accuracy and ultimately enabling efficient utilization of the lidar equipment.

[0058] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] like Figure 1 and Figure 2 As shown, Figure 1 A schematic diagram of a scanning area provided by the lidar device in the embodiments of this application; Figure 2 for Figure 1 A schematic diagram of the LiDAR device within the scanning area of ​​the XOZ plane.

[0060] like Figure 1 and Figure 2 The lidar device 10 shown, based on the three-dimensional rectangular coordinate system O-XYZ established at the center of the lidar device 10, performs a 360° all-around scan in the XOY plane, while scanning within a certain pitch angle range in the Z-axis direction.

[0061] In practical applications, the region of interest that the lidar device 10 needs to scan and measure in the XOY plane often does not cover a 360° omnidirectional scanning area. Therefore, reflective elements can be placed within the scanning area of ​​the lidar device 10.

[0062] For ease of understanding and explanation, in the subsequent embodiments of this application, the lidar equipment involved is mainly based on... Figure 1 and Figure 2 The lidar device 10 shown is used as an example for introduction. Because Figure 1 and Figure 2 The lidar device 10 shown has a large scanning area in the XOY plane. Therefore, the subsequent embodiments in this application are mainly described with the reflective element set in the XOY plane, that is, the reflective surface is parallel to the Z-axis direction. In the subsequent embodiments, the lidar device 10 scans 360° in the XOY plane, and the reflective surface of the reflective element is parallel to the Z-axis direction. These prerequisites will not be repeated.

[0063] Of course, it is understood that the calibration method for the relative positional relationship between the lidar device and the reflective element in this application is not limited to the example in which the lidar device performs a 360° omnidirectional scan in the XOY plane and the reflective surface of the reflective element is parallel to the Z-axis direction. Following the same principle as the technical solution provided in this application, the calibration of the relative positional relationship between the lidar device and the reflective element in other similar layout examples with different reflective elements can also be achieved.

[0064] like Figure 3 As shown, Figure 3 That is to be in Figure 1 A schematic diagram of an optical path in which reflective elements are set in the scanning area of ​​the XOY plane of a lidar device.

[0065] exist Figure 3 In the example shown, the reflective element 20 includes two mutually perpendicular first plane mirrors 21 and second plane mirrors 22. The first and second plane mirrors 21 and 22 are symmetrical about the X-axis, and their reflective surfaces are parallel to the Z-axis. Furthermore, the two outer sides of the first and second plane mirrors 21 and 22, which are facing away from each other, are located on the Y-axis. From the XOY scanning orientation, part of the laser light output by the lidar device 10 is reflected once by one plane mirror and then output in a direction away from the reflective element. Another portion of the laser light is reflected twice by the two plane mirrors and then output in a direction away from the reflective element. Ultimately, this reduces the scanning area of ​​the lidar device 10 in the XOY plane to nearly half, allowing the laser light to radiate more concentratedly and densely within the reduced scanning area.

[0066] Understandably, in practical applications, Figure 3 The reflective element 20 shown should obviously be positioned on the side of the lidar device 10 away from the target object to be scanned. That is to say, in this application, the laser light output from the lidar device 10 towards the side away from the target object is deflected, so that this portion of the light, after reflection, can also be scanned and output towards the target object. This ensures that the laser light output from the lidar device 10 towards the side away from the target object can also be utilized, and also increases the density of the laser light output from the lidar device 10 towards the target object, thereby improving the accuracy of scanning and measuring the target object.

[0067] For ease of understanding, in the following... Figure 5 , Figure 6 In the embodiment shown, both the first plane mirror 21 and the second plane mirror 22 are... Figure 3 The layout of the reflective elements 20 is exactly the same, and will be discussed later. Figure 5 and Figure 6 In the embodiments described above, the position setting conditions satisfied by the first plane mirror 21 and the second plane mirror 22 will not be repeated.

[0068] Of course, it is understandable that in practical applications where the laser radar device 10 works in conjunction with the device to achieve a more concentrated laser beam scanning range, it is not necessarily required that the following conditions be met. Figure 3 The layout of the reflective elements 20 shown in this application, and the calibration method in this application, are not limited to implementing... Figure 3 The parameter calibration of the relative positional relationship between the reflective element 20 and the lidar device 10 shown.

[0069] Based on the above discussion, this application further provides a lidar calibration method for calibrating parameters between lidar equipment and reflective elements, thereby providing reliable data support for the use of lidar equipment in conjunction with reflective elements and ensuring the scanning accuracy of lidar equipment.

[0070] like Figure 4 As shown, a lidar calibration method provided in this application may include:

[0071] M1: Based on the relative position between the reflective element and the lidar device, the point cloud data collected by the lidar device is divided into reflective virtual image point cloud and non-reflective point cloud; among them, the reflective virtual image point cloud is collected by scanning the laser light reflected by the reflective element.

[0072] For ease of understanding, Figure 3 Taking the optical path structure shown as an example, in the scanning area of ​​the 360° omnidirectional scan in the XOY plane, half of the laser light rays near the reflector 20 are reflected and then scanned out, while the other half of the laser light rays output to the side away from the reflector 20 are not reflected. However, regardless of whether the laser light rays are reflected by the reflector 20 or not, corresponding sampling points can be scanned and collected. All sampling points together constitute the point cloud data collected by the lidar device 10. Therefore, the sampling points obtained by scanning the laser light rays reflected by the reflector in the point cloud data can be classified as reflected virtual image point clouds, while the sampling points not reflected by the reflector can be classified as non-reflective point clouds.

[0073] Furthermore, according to the basic working principle of lidar, a laser beam is emitted by a transmitter and travels in a straight line. After being reflected by the surface of the target object, it returns along the original path and is received by a receiver. Based on the time difference between the emission and reception of the laser beam, the speed of light traveling in a straight line in space, and the azimuth angle of the laser beam emission, the coordinates of a point on the target object's surface reflecting the laser beam can be determined. The point corresponding to this coordinate value is a sampling point. The laser scanning device scans the laser beams in different directions in space and collects the sampling points, which together form point cloud data.

[0074] This demonstrates that lidar equipment, during the point cloud data acquisition process, did not consider the potential for multiple deflections in the transmission path of laser light after reflection by reflective elements. Therefore... Figure 5As shown, for position point S1 in space, when a laser beam is reflected once and then incident on position point S1, and then returns along the original path to be received by the receiver, since the reflection process was not considered when determining the position of position point S1 in space, the coordinates of the finally determined sampling point are the coordinates of position point S1'. Obviously, position point S1' is the virtual image position of position point S1. For example, for position point S2 in space, when a laser beam is reflected twice and incident on position point S2, the coordinates of the sampling point determined by this laser beam should be the coordinates of position point S2'. Obviously, position point S2' is the virtual image position corresponding to position point S2 after the two reflections, where position point S21' is the virtual image position of the first reflection and position S22' is the virtual image position of the second reflection.

[0075] Therefore, it is evident that when lidar equipment and reflective elements work together, the coordinates of the sampling points determined by the laser light reflected by the reflective elements are not the actual coordinates of the target object's surface. Based on this, it is necessary to transform the coordinates of these sampling points to determine the coordinates of the actual object points (e.g., position points S1 and S2). To achieve this, it is first necessary to distinguish between reflected virtual image point clouds and non-reflected point clouds in the point cloud data.

[0076] Optionally, dividing the point cloud data into reflective virtual image point clouds and non-reflective point clouds may include:

[0077] M11: Based on the relative position between the reflective element and the lidar device, the scanning area of ​​the lidar device is divided into a reflective scanning angle area and a non-reflective scanning angle area;

[0078] M12: Divides point cloud data acquired through the reflective scanning angle region into reflective virtual image point cloud, and divides point cloud data acquired through the non-reflective scanning angle region into non-reflective point cloud.

[0079] Based on the relative positional relationship between the reflective element and the lidar device in this embodiment, it is clear that within the scanning range of the lidar device in the XOY plane, it can be roughly determined which part of the angled output light will be reflected by the reflective element. Obviously, the sampling points corresponding to the laser light output at that angle are classified as reflected virtual image point clouds, while the sampling points obtained by scanning the laser light that is directly output towards the target without being reflected by the reflective element are classified as non-reflective point clouds.

[0080] Further optionally, taking a reflective element including a first plane mirror and a second plane mirror as an example; wherein one end of the first plane mirror and the second plane mirror are spliced ​​together, and the angle between the reflective surfaces is within a set angle range.

[0081] Based on this, the process of determining the reflection scanning angle region and the non-reflection scanning angle region in M11 above can include:

[0082] M1111: Based on the relative position between the reflective element and the lidar device, determine the intermediate scanning direction, the first boundary direction, and the second boundary direction for the lidar device to scan the splicing position between the first and second planar reflectors and the mutually opposing side positions.

[0083] M1112: The scanning angle region between the first boundary direction and the middle scanning direction, closer to the lidar device, is taken as the initial first reflection angle region;

[0084] M1113: The scanning angle region between the intermediate scanning direction and the second boundary direction is used as the initial second reflection angle region;

[0085] M1114: Reduce the initial first reflection angle region and the initial second reflection angle region from both ends of the boundary by a fine-tuning angle to determine the first reflection angle region and the second reflection angle region;

[0086] M1115: The first reflection angle region and the second reflection angle region are used as the reflection scanning angle region, and the scanning angle region on the side away from the lidar device between the first boundary direction and the second boundary direction is used as the non-reflection scanning angle region.

[0087] by Figure 6 For example, in Figure 6 In the illustrated embodiment, the scanning directions of the laser light output by the lidar device 10 to the outer edges of the first planar reflector 21 and the second planar reflector 22 are the positive and negative Y-axis directions, respectively. The scanning direction of the laser light output to the junction between the first and second planar reflectors is the positive X-axis direction. Therefore, the 90° angle region between the positive Y-axis and the positive X-axis is the scanning area where the lidar device 10 outputs laser light to the first planar reflector 21, i.e., the initial first reflection angle region. The 90° angle region between the positive X-axis and the positive Y-axis is the scanning area where the lidar device 10 outputs laser light to the second planar reflector 22, i.e., the initial second reflection angle region. The scanning range within the 180° range from the positive Y-axis to the negative X-axis and back to the negative Y-axis is clearly the non-reflective scanning angle region.

[0088] Of course, in Figure 6 In the illustrated embodiment, the angle between the reflective surfaces of the first plane mirror 21 and the second plane mirror 22 is 90° as an example. In practical applications, the angle between the first plane mirror 21 and the second plane mirror 22 is not necessarily 90°. Figure 7 As shown, Figure 7 The first reflecting plane mirror 21 and the second reflecting plane mirror 22 are also spliced ​​together, symmetrical about the X-axis, and their reflective surfaces are parallel to the Z-axis, but... Figure 6 The embodiment shown differs from the one described above in that... Figure 7 In the illustrated embodiment, the angle between the first reflecting plane mirror 21 and the second reflecting plane mirror 22 is greater than 180°. Figure 7 The embodiments shown are in accordance with the above. Figure 6 A similar method can be used to determine the initial first reflection angle region, which is... Figure 7 The angle a0 in the middle, and the initial second reflection angle region is the... Figure 7 Angle b0 in the XOY plane; while the angle region c outside of angles a0 and b0 in the XOY plane is the non-reflection angle region.

[0089] Based on this, it is further considered that the optical path of the laser beam output by the lidar device 10 scanning the edge regions of the first planar reflector 21 and the second planar reflector 22 is often quite complex, especially as... Figure 6 In the embodiment shown, in the area where the first plane mirror 21 and the second plane mirror 22 are joined together, the transmission path of the laser light may be reflected and deflected more than twice, resulting in a large error.

[0090] Therefore, in this embodiment, the boundaries of the initial first reflection angle region and the initial second reflection angle region can be further fine-tuned. A set fine-tuning angle can be selected, and the initial first reflection angle region can be reduced from both sides by the set fine-tuning angle. That is, while keeping the center angle of the initial first reflection angle region unchanged, the angle region of the entire initial first reflection angle region is reduced by half of the set fine-tuning angle, thus obtaining the adjusted first reflection angle region. For example... Figure 7 As shown, the two sides of the initial first reflection angle region a0 are finely adjusted according to the set angle. By adjusting and reducing the zoom level, the first reflection angle region can be obtained, that is... Figure 7 Angle a in the middle.

[0091] Similarly, for the initial second reflection angle region b0, the angle can be finely adjusted according to the set parameters while keeping the center angle unchanged. Adjustments are made to both sides of the boundary to reduce the size of the area and obtain the second reflection angle region b.

[0092] Based on this, and according to the principle of optical reflection, when a reflective element contains two or more plane mirrors spliced ​​together, if the angle between the reflective surfaces of two adjacent plane mirrors is less than 180°, some of the laser light output by the lidar will inevitably be reflected twice or more between the two plane mirrors.

[0093] Therefore, in another optional embodiment, when the reflective element includes a first plane mirror and a second plane mirror that are spliced ​​together and symmetrically arranged about the X-axis, and the angle between the reflective surfaces of the first plane mirror and the second plane mirror is less than 180 degrees, the process of determining the reflection scanning angle region may further include:

[0094] M1121: Based on the relative position between the first planar reflector and the lidar, the first reflection angle region is divided into a first primary reflection angle region and a first secondary reflection angle region;

[0095] M1122: Based on the relative position between the second plane mirror and the lidar, the second reflection angle region is divided into a second primary reflection angle region and a second secondary reflection angle region;

[0096] M1123: The first primary reflection angle region, the first secondary reflection angle region, the second primary reflection angle region, and the second secondary reflection angle region are collectively used as the reflection scanning angle region.

[0097] Reference Figure 6 The method for determining the first initial reflection angle region and the second initial reflection angle region in this embodiment is similar to that in the previous embodiment, and will not be repeated here.

[0098] Based on this, in the first initial reflection angle region, part of the laser light emitted by the lidar device is reflected once by the first plane mirror before being emitted towards the target, while another part of the laser light is reflected a second time by the second plane mirror before being emitted towards the target.

[0099] The parameters are as described above. Figure 5 In the illustrated embodiment, the positional relationship between the sampling points and the object points obtained from a single reflection of the laser beam and from a double reflection is also different. Therefore, in this embodiment, the first initial reflection angle region can be further divided into scanning regions that have been reflected once and those that have been reflected twice, in order to distinguish between sampling points obtained from laser beams that have been reflected once and those obtained from laser beams that have been reflected twice.

[0100] Based on the principle of optical reflection, it is known that for the first plane mirror, the angular boundary between the laser light output by the lidar device scanning the first plane mirror and the laser light that will be reflected only once and then reflected a second time by the second plane mirror is the direction perpendicular to the first plane mirror.

[0101] For example Figure 6Taking the illustrated embodiment as an example, the direction in which the lidar device 10 vertically scans and outputs the laser light to the first planar reflector 21 is defined as the vector. The direction of the laser beam output by the lidar device 10 in a perpendicular scan of the second planar reflector 22 is the vector direction. The direction.

[0102] Obviously, the lidar device has a 10-axis vector control. Laser rays radiated from the region at the angle between the Y-axis and the positive Y-axis are reflected only once by the first planar mirror 21, thus the vector The angle region between the positive Y-axis and the positive Y-axis is the first reflection angle region in this embodiment; however, the lidar device has a 10-axis vector. The laser beam radiated from the region at the angle between the X-axis and the positive X-axis will be reflected once by the first plane mirror 21, and then reflected a second time by the second plane mirror 22. Thus, this vector... The angle region between the X-axis and the positive X-axis is also the first and second reflection angle region in this embodiment.

[0103] Similarly, for the 10-axis vector control of lidar devices... The region between the laser beam and the negative Y-axis is the second reflection angle region, vector The angular region between the positive X-axis and the positive X-axis is the second reflection angle region.

[0104] Based on this, it is further considered that there is a large probability of error in the sampling points collected by the lidar device in the area where the first and second plane mirrors are spliced ​​together, the two outer side areas where the first and second plane mirrors are facing away from each other, and the areas where the lidar device outputs laser light to the first and second plane mirrors at near-vertical angles.

[0105] Therefore, refer to Figure 6 In this embodiment, vector The first reflection angle region formed by the angle between the positive Y-axis and the positive Y-axis, and the vector The first and second reflection angle regions formed by the angle between the X-axis and the positive X-axis, and the vector The second reflection angle region formed by the angle between the Y-axis and the negative Y-axis, and the vector The second and third reflection angle regions formed by the angle between the X-axis and the positive X-axis can be adjusted by reducing the angle region according to a set fine-tuning angle. The adjustment method is similar to the above embodiment, that is, keeping the center angle of each angle region unchanged, reducing the set fine-tuning angle on both sides of the angle region from the boundary on both sides, so that the entire angle region is reduced by half of the set fine-tuning angle; for example Figure 6As shown, the adjusted first reflection angle region, first and second reflection angle region, second reflection angle region, and second reflection angle region correspond to respectively Figure 6 Angles a1, angle a2, angle b1, and angle b2 are shown in the figure.

[0106] Based on the above discussion, after dividing the reflection scanning angle area corresponding to the lidar device into four angle areas in sequence—the first primary reflection angle area, the first secondary reflection angle area, the second primary reflection angle area, and the second secondary reflection angle area—the point cloud data output through the first primary reflection angle area can be further divided into the first primary reflection virtual image point cloud, the point cloud data output through the first secondary reflection angle area into the first secondary reflection virtual image point cloud, the point cloud data output through the second primary reflection angle area into the second primary reflection virtual image point cloud, and the point cloud data output through the second secondary reflection angle area into the second secondary reflection virtual image point cloud.

[0107] M2: Based on the principle of optical reflection and the initial reflective surface equation of the reflective element, coordinate transformation is performed on each virtual image sampling point in the reflected virtual image point cloud to obtain the real image sampling point.

[0108] As described above, the coordinates of the sampling points collected after the laser light output by the lidar device in this application is essentially the virtual image coordinates of the real position points on the surface of the target object; however, to measure the target object, it is obviously necessary to know the coordinates of the real position points on the surface of the target object; that is, it is necessary to transform the virtual image coordinates according to the relative positional relationship between the reflector and the lidar.

[0109] In a three-dimensional Cartesian coordinate system with the center of the lidar device as the origin, the reflective surface of the reflective element can be represented by different surface shapes. For example, if the reflective surface of the reflective element is a plane mirror, it can be represented by a plane equation in the three-dimensional Cartesian coordinate system. If the reflective element is a parabolic surface, a partial sphere, a cylindrical surface, etc., its reflective surface can be represented by the parabolic surface equation, spherical surface equation, and cylindrical surface equation in the three-dimensional Cartesian coordinate system, respectively. After the equation satisfied by the reflective surface of the reflective element is determined, the transformation relationship between each virtual image sampling point and the corresponding real position point coordinates on the target surface can be determined, that is, the transformation relationship between virtual image sampling points and real image sampling points in this embodiment.

[0110] Optionally, in embodiments where the reflective element includes a first and a second plane mirror that are spliced ​​together and symmetrical about the X-axis, the process of performing coordinate transformation on each virtual image sampling point in the reflected virtual image point cloud to obtain real image sampling points may include:

[0111] M21: Based on the relative positional relationship between the lidar device and the first planar reflector, according to the aforementioned reflective surface equation model... Construct the initial equation of the first reflective surface of the first planar reflector in a three-dimensional rectangular coordinate system; wherein the origin of the three-dimensional rectangular coordinate system coincides with the center of the lidar device; wherein, All are coefficient parameters;

[0112] M22: Based on the relative positional relationship between the lidar device and the second plane mirror, according to the reflective surface equation model Construct the initial equation of the second reflective surface of the second plane mirror in the three-dimensional rectangular coordinate system;

[0113] M23: Initial first reflective surface equation, initial second reflective surface equation, and reflection conversion matrix model; construct the first and second reflection conversion matrices; wherein, the reflection conversion matrix model is predetermined based on the principle of optical reflection;

[0114] M24: The first primary reflection virtual image point cloud is transformed using the first reflection transformation matrix. The first and second secondary reflection virtual image point clouds are then transformed using the first and second reflection transformation matrices in sequence. The second primary reflection virtual image point cloud is then transformed using the second reflection transformation matrix. The second and second secondary reflection virtual image point clouds are then transformed using the second and first reflection transformation matrices in sequence to obtain real image sampling points.

[0115] As mentioned above, since both the first and second plane mirrors have planar reflective surfaces, their reflective surfaces both satisfy... This form of equation, namely the initial first reflector equation and the initial second reflector equation, can both be expressed as follows: This reflective surface equation model takes the form of, where the initial first reflective surface equation and the initial second reflective surface equation are derived from the different relative positions of the first and second plane mirrors with the lidar device. The coefficient parameters should be different for each. Based on this, the initial equation of the first reflective surface can be written as: The initial equation for the second reflective surface is written as: Obviously, based on the pre-defined relative positions of the lidar device and the first and second planar reflectors, it is theoretically possible to determine... Parameters such as these are important; however, in the actual assembly and installation of lidar equipment and reflective elements, the relative positional relationship between the two cannot be guaranteed to be absolutely accurate. For lidar equipment scanning and measurement, even a small deviation in the relative position between the lidar and the reflective element can greatly affect the accuracy of the lidar.

[0116] Therefore, for embodiments where the reflective elements are two plane mirrors, In other words, these are the parameters that need to be calibrated in this embodiment to characterize the relative positional relationship between the lidar device and the reflective element. Similarly, in embodiments where the reflective element is another type of mirror, the required calibration parameters are also the coefficients in the reflective surface equation; this will not be elaborated upon further in this application.

[0117] Furthermore, in this embodiment, the theoretical relative positions of the lidar device, the first planar reflector, and the second planar reflector can be used to determine... The initial values ​​of the coefficient parameters are used as a basis for subsequent optimization and iterative updates of the above coefficient parameters, so that more accurate coefficient parameters can be determined, which are the parameters that need to be calibrated in this application.

[0118] Based on this, after determining the initial first reflective surface equation and the initial second reflective surface equation respectively, the coordinate transformation relationship matrix between the real image point and the virtual image point of the light rays reflected by the first plane mirror and the second plane mirror can be determined according to the pre-determined reflection transformation matrix.

[0119] The reflective conversion matrix model in this embodiment is pre-determined based on the principle of planar reflectivity. The process of determining this reflective conversion matrix model may include:

[0120] M201: According to the equation of the reflective surface... Normal vector of the mirror The unit normal vector is determined as .

[0121] S202: Introducing the first translation matrix that represents the translation of the origin of the coordinate system to the reflective surface is: The second translation matrix, which represents the translation of points on the reflective surface to the origin of the coordinate system, is: .

[0122] M203: Construct the Haushold reflection transformation matrix as follows ;in, It is a unit vector.

[0123] M204: Based on the Haushold reflection transformation matrix, the planar reflection matrix of the mirror is constructed as follows: .

[0124] M205: Based on the first translation matrix, the second translation matrix, and the reflection matrix, perform matrix multiplication in the order of translation, reflection, and finally translation back to the original coordinate system to obtain the reflection transformation matrix model. .

[0125] Therefore, the reflection transformation matrix model determined in this embodiment is:

[0126] .

[0127] Based on the above discussion of the reflection conversion matrix model, and combined with the initial equation of the first reflective surface, the first reflection conversion matrix can be determined as follows:

[0128] ;

[0129] in, .

[0130] Similarly, by combining the initial equation of the first reflective surface, the first reflection conversion matrix can be determined as follows:

[0131] ;

[0132] in, .

[0133] Based on this, for each virtual image sampling point in the first reflection virtual image point cloud, according to the aforementioned first reflection transformation matrix, according to... Perform the conversion; among which, The coordinates of the virtual image sampling point in the first reflection virtual image point cloud are given. for The coordinates of the corresponding real image sampling point.

[0134] The virtual image sampling points in the first and second reflection virtual image point clouds are then processed according to the first and second reflection transformation matrices described above. Perform the conversion; The coordinates of the virtual image sampling points in the first and second reflection virtual image point clouds are given. for The coordinates of the corresponding real image sampling point.

[0135] For each virtual image sampling point in the second reflection virtual image point cloud, according to the aforementioned second reflection transformation matrix, according to... Perform the conversion; among which, The coordinates of the virtual image sampling point in the second reflection virtual image point cloud are given. for The coordinates of the corresponding real image sampling point.

[0136] The virtual image sampling points in the second and second reflection virtual image point cloud are then processed according to the first and second reflection transformation matrices described above. Perform the conversion; The coordinates of the virtual image sampling points in the first and second reflection virtual image point clouds are given. for The coordinates of the corresponding real image sampling point.

[0137] M3: Search for the nearest neighboring sampling point to the real image sampling point in the non-reflection point cloud, and construct the objective function based on the deviation between the real image sampling point and the corresponding neighboring sampling point.

[0138] As mentioned above, after a portion of the laser light output by the lidar device is emitted by the reflective element, it can be concentrated and radiated onto the same area scanned by the unreflected laser light. Therefore, the real image sampling points obtained by converting the virtual image sampling points collected by the reflected laser light should be the position points on the same surface of the target object corresponding to each sampling point in the non-reflective point cloud.

[0139] Based on this, we can search for neighboring sampling points of each real image sampling point in the non-reflection point cloud. Specifically, we can search for the sampling point with the smallest Euclidean distance between it and the real image sampling point as the neighboring sampling point. Of course, the Euclidean distance between the neighboring sampling point and the corresponding real image sampling point should not be greater than a set distance threshold.

[0140] An objective function can be constructed based on the deviation between each real image sampling point and its corresponding neighboring sampling points; specifically, the objective function can be constructed based on the distance deviation between each real image sampling point and its corresponding neighboring sampling points. ;in, This represents the parameters to be optimized in the initial reflective surface equation. This represents the total number of real image sampling points. Indicates in the parameter Under the conditions One real image sampling point; Indicates the first The neighboring sampling points corresponding to each real image sampling point.

[0141] It is understood that in this embodiment... That is, the parameters that need to be optimized, for example, in embodiments that correspond to reflective elements including a first plane mirror and a second plane mirror. That is to include .

[0142] It is understood that the objective function in this embodiment is not limited to the form described above. For example, in another optional embodiment, surface fitting can be performed on real image sampling points with neighboring sampling points to obtain the equation of the real image plane where each real image sampling point is located. Centroid point calculation can also be performed on each real image sampling point to obtain the real image centroid. Then, surface fitting can be performed on the neighboring sampling points corresponding to each real image sampling point to obtain the equation of the neighboring point plane where each neighboring sampling point is located. Centroid point calculation can also be performed on each neighboring sampling point to obtain the neighboring centroid. Then, the objective function is constructed by weighted summing the normal force angle deviation between the realized plane equation and the neighboring plane equation and the distance deviation between the real image centroid and the neighboring centroid. This can also achieve the technical solution of this application.

[0143] M4: Optimize and iterate the parameters in the initial reflective surface equation to obtain calibration parameters that make the objective function less than the target threshold.

[0144] After determining the objective function, the parameters in the initial reflective surface equation can be repeatedly optimized and iterated until the function value of the objective function is less than the target threshold or the number of optimization iterations reaches the upper limit, thus obtaining the calibration parameters.

[0145] In summary, this application incorporates a reflective element in the scanning area of ​​the lidar device. By reflecting and deflecting the laser light from non-interest areas using the reflective element, the laser light from the lidar device can be more concentrated and radiated towards the area of ​​interest, thereby increasing the laser scanning density in the area of ​​interest and improving scanning accuracy, thus enhancing the efficient utilization of the lidar device.

[0146] Based on this, the laser light reflected by the reflective element needs to be based on the precise relative positional relationship between the reflective element and the lidar device to achieve accurate scanning measurement. Therefore, this application divides the point cloud data collected by the lidar device into reflected virtual image point clouds and non-reflective point clouds. Based on the principle of optical reflection, the reflected virtual image point clouds are transformed to obtain real image sampling points. Finally, based on the deviation between the real image sampling points and the non-reflective point clouds in the scanning results of the same scanning area, the initial reflective surface equation characterizing the relative position between the reflective element and the lidar device is optimized and iterated. Ultimately, the parameters characterizing the relative position between the reflective element and the lidar device are accurately calibrated, providing possible data basis for the lidar device to combine with the reflective element to accurately scan and measure the region of interest.

[0147] Based on the above discussion, in another optional embodiment of this application, such as Figure 8 As shown, the reflective element 20 disposed in the scanning area of ​​the lidar device 10 includes:

[0148] The first plane mirror 21 and the second plane mirror 22 are spliced ​​together with the reflective surface angle less than 180 degrees, and the third plane mirror 23 and the fourth plane mirror 24 are respectively set on both sides of the first plane mirror 21 and the second plane mirror 22; the first plane mirror 21 and the third plane mirror 23, as well as the second plane mirror 22 and the fourth plane mirror 24 are spliced ​​together with the reflective surface angle greater than 180 degrees and less than 270 degrees.

[0149] Furthermore, the first plane mirror 21 and the second plane mirror 22, as well as the third plane mirror 23 and the fourth plane mirror 24, are all symmetrical about the lidar device 10.

[0150] Accordingly, the lidar calibration method may include:

[0151] Step 1): Based on the relative positions of the first, second, third, and fourth plane mirrors and the lidar, determine the non-reflective scanning angle region, and the first, second, third, and fourth reflection angle regions within the reflective scanning angle region.

[0152] Step 2): Divide the point cloud data acquired through the first reflection angle region, the second reflection angle region, the third reflection angle region, and the fourth reflection angle region into the first reflection virtual image point cloud, the second reflection virtual image point cloud, the third reflection virtual image point cloud, and the fourth reflection virtual image point cloud; divide the point cloud data acquired through the non-reflection scanning angle region into the non-reflection point cloud.

[0153] Step 3) Based on the principle of optical reflection and the initial first reflective surface equation, initial second reflective surface equation, initial third reflective surface equation, and initial second reflective surface equation corresponding to the first, second, third, and fourth plane reflective mirrors, respectively, coordinate transformation is performed on the first, second, third, and fourth reflection virtual image point clouds to obtain real image sampling points.

[0154] Step 4) Search for the nearest neighboring sampling point in the non-reflection point cloud that is closest to the real image sampling point, and construct the objective function based on the deviation between the real image sampling point and the corresponding neighboring sampling point.

[0155] Step 5) Optimize and iterate the parameters in the initial reflective surface equation to obtain calibration parameters that make the objective function less than the target threshold.

[0156] As can be seen from this embodiment, the reflective element in this application can be composed of two or more plane mirrors spliced ​​together, and even if the reflective element includes four plane mirrors, it is not necessarily as... Figure 8 The settings shown are as follows: Figure 9 As shown, the angle between the reflective surfaces of the first plane mirror 21 and the second plane mirror 22 can also be greater than 180°, while the angle between the reflective surfaces of the first plane mirror 21 and the third plane mirror 23, and between the second plane mirror 22 and the fourth plane mirror 24 can all be less than 180°, which can also achieve the technical solution of this application.

[0157] Further optionally, the reflective surface angle between the first and second plane mirrors, the reflective surface angle between the first and third plane mirrors, and the reflective surface angle between the third and fourth plane mirrors are all adjustable by a drive motor;

[0158] Accordingly, the process of determining the relative position between the reflective element and the lidar device may include:

[0159] Based on the driving parameters of the drive motors for the first, second, third, and fourth plane mirrors, the relative positions of the first, second, third, and fourth plane mirrors to the lidar device are determined.

[0160] In this embodiment, the first and second planar reflectors rotate relative to each other using their joint connection point as a fulcrum, thereby changing the reflective surfaces between them. Similarly, similar driving adjustments can be achieved between the first and third planar reflectors, and between the second and fourth planar reflectors. Therefore, in practical applications, the lidar device of this application can reasonably adjust the angles between the reflective surfaces of the first and second planar reflectors, the first and third planar reflectors, and the second and fourth planar reflectors based on actual needs, thereby better meeting actual measurement requirements. Furthermore, after each adjustment, the parameters between the reflective elements and the lidar device can be calibrated according to the calibration method described in this application, thereby achieving accurate measurements by the lidar device.

[0161] Of course, it is understood that the reflective elements in this application are not limited to multiple flexibly adjustable reflective elements, but can also be fixed planar or curved reflective mirrors; and the relative position between the lidar device and the reflective element is adjustable. After each adjustment of the relative position between the lidar and the reflective element, the parameters between the reflective element and the lidar device can be calibrated according to the calibration method of this application, thereby achieving accurate measurement of the lidar device. This application will not list all examples of this.

[0162] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that the elements inherent in a process, method, article, or apparatus that includes a list of elements are included. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Additionally, portions of the technical solutions provided in the embodiments of this application that are consistent with the implementation principles of corresponding technical solutions in the prior art have not been described in detail to avoid excessive elaboration.

[0163] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. It should be noted that those skilled in the art can make several improvements and modifications to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the protection scope of the present invention.

Claims

1. A lidar calibration method, characterized in that, A reflective element is provided within the scanning area of ​​a lidar device to reflect laser light in a portion of the area; the calibration method includes: Based on the relative position between the reflective element and the lidar device, the point cloud data collected by the lidar device is divided into reflected virtual image point cloud and non-reflective point cloud; wherein, the reflected virtual image point cloud is collected by scanning the laser light reflected by the reflective element; Based on the principle of optical reflection and the initial reflective surface equation of the reflective element, coordinate transformation is performed on each virtual image sampling point in the reflected virtual image point cloud to obtain real image sampling points; Search for the nearest neighboring sampling point to the real image sampling point in the non-reflection point cloud, and construct an objective function based on the deviation between the real image sampling point and the corresponding neighboring sampling point; The parameters in the initial reflective surface equation are optimized and iterated to obtain calibration parameters that make the objective function less than the target threshold.

2. The lidar calibration method as described in claim 1, characterized in that, Based on the relative position between the reflective element and the lidar device, the point cloud data collected by the lidar device is divided into reflective virtual image point clouds and non-reflective point clouds, including: Based on the relative position between the reflective element and the lidar device, the scanning area of ​​the lidar device is divided into a reflective scanning angle area and a non-reflective scanning angle area; The point cloud data acquired through the reflection scanning angle region is divided into the reflection virtual image point cloud, and the point cloud data acquired through the non-reflection scanning angle region is divided into the non-reflection point cloud.

3. The lidar calibration method as described in claim 2, characterized in that, The reflective element includes a first planar reflector and a second planar reflector; wherein one end of the first planar reflector and the second planar reflector are spliced ​​together, and the angle between the reflective surfaces is within a set angle range; Based on the relative position between the reflective element and the lidar device, the scanning area of ​​the lidar device is divided into a reflective scanning angle region and a non-reflective scanning angle region, including: Based on the relative position between the reflective element and the lidar device, the middle scanning direction, the first boundary direction, and the second boundary direction of the lidar device for scanning the splicing position between the first planar reflector and the second planar reflector and the mutually opposite side positions are determined; The scanning angle region between the first boundary direction and the intermediate scanning direction, closer to the lidar device, is taken as the initial first reflection angle region. The scanning angle region between the intermediate scanning direction and the second boundary direction is used as the initial second reflection angle region; The initial first reflection angle region and the initial second reflection angle region are both reduced from their two end boundaries by a fine-tuning angle to determine the first reflection angle region and the second reflection angle region. The first reflection angle region and the second reflection angle region are used as the reflection scanning angle region, and the scanning angle region on the side away from the lidar device between the first boundary direction and the second boundary direction is used as the non-reflection scanning angle region.

4. The lidar calibration method as described in claim 3, characterized in that, When the first planar reflector and the second planar reflector are symmetrically arranged about the lidar device, and the angle between the reflective surfaces is less than 180 degrees, then the first reflection angle region and the second reflection angle region are used as the reflection scanning angle region, including: Based on the relative position between the first planar reflector and the lidar, the first reflection angle region is divided into a first primary reflection angle region and a first secondary reflection angle region; Based on the relative position between the second planar reflector and the lidar, the second reflection angle region is divided into a second primary reflection angle region and a second secondary reflection angle region. The first primary reflection angle region, the first secondary reflection angle region, the second primary reflection angle region, and the second secondary reflection angle region are collectively referred to as the reflection scanning angle region; Accordingly, the point cloud data acquired through the reflection scanning angle region is divided into the reflection virtual image point cloud, including: The point cloud data output through the first primary reflection angle region is divided into a first primary reflection virtual image point cloud, the point cloud data output through the first secondary reflection angle region is divided into a first secondary reflection virtual image point cloud, the point cloud data output through the second primary reflection angle region is divided into a second primary reflection virtual image point cloud, and the point cloud data output through the second secondary reflection angle region is divided into a second secondary reflection virtual image point cloud.

5. The lidar calibration method as described in claim 4, characterized in that, Based on the principle of optical reflection and the initial reflective surface equation of the reflective element, coordinate transformation is performed on each virtual image sampling point in the reflected virtual image point cloud to obtain real image sampling points, including: Based on the relative positional relationship between the lidar device and the first planar reflector, and according to the reflective surface equation model... The initial equation of the first reflective surface of the first planar reflector in a three-dimensional Cartesian coordinate system is constructed; wherein the origin of the three-dimensional Cartesian coordinate system coincides with the center of the lidar device; wherein, All are coefficient parameters; Based on the relative positional relationship between the lidar device and the second planar reflector, and according to the reflective surface equation model... Construct the initial equation of the second reflective surface of the second planar reflector in the three-dimensional rectangular coordinate system; Based on the initial first reflective surface equation, the initial second reflective surface equation, and the reflection conversion matrix model, a first reflection conversion matrix and a second reflection conversion matrix are constructed; wherein, the reflection conversion matrix model is determined in advance based on the principle of optical reflection; The first primary reflection virtual image point cloud is transformed using the first reflection transformation matrix. The first secondary reflection virtual image point cloud is transformed using the first reflection transformation matrix and the second reflection transformation matrix in sequence. The second primary reflection virtual image point cloud is transformed using the second reflection transformation matrix in sequence. The second secondary reflection virtual image point cloud is transformed using the second reflection transformation matrix and the first reflection transformation matrix in sequence to obtain real image sampling points.

6. The lidar calibration method as described in claim 5, characterized in that, The predetermined reflection transformation matrix model is as follows: ;in, , , And the unit normal vector .

7. The lidar calibration method as described in claim 2, characterized in that, The reflective element includes a first plane mirror and a second plane mirror spliced ​​together with a reflective surface angle of less than 180 degrees, and a third plane mirror and a fourth plane mirror respectively disposed on both sides of the first plane mirror and the second plane mirror; the first plane mirror and the third plane mirror, as well as the second plane mirror and the fourth plane mirror, are spliced ​​together with a reflective surface angle of greater than 180 degrees and less than 270 degrees. Accordingly, based on the relative position between the reflective element and the lidar device, the scanning area of ​​the lidar device is divided into a reflective scanning angle region and a non-reflective scanning angle region, including: Based on the relative positions of the first, second, third, and fourth planar reflectors and the lidar, the non-reflective scanning angle region, as well as the first, second, third, and fourth reflection angle regions within the reflective scanning angle region, are determined.

8. The lidar calibration method as described in claim 7, characterized in that, The reflective surface angles between the first and second plane mirrors, between the first and third plane mirrors, and between the third and fourth plane mirrors are all adjustable via a drive motor. Accordingly, the process of determining the relative position between the reflective element and the lidar device, before separating the reflective virtual image point cloud and the non-reflective point cloud from the point cloud data collected by the lidar device, further includes: Based on the driving parameters of the drive motor for the first, second, third, and fourth planar reflectors, the relative positions of the first, second, third, and fourth planar reflectors and the lidar device are determined.

9. The lidar calibration method as described in claim 1, characterized in that, Searching for the nearest neighbor sampling point in the non-reflective point cloud that is closest to the real image sampling point includes: For each real image sampling point, a sampling point with the smallest Euclidean distance to the real image sampling point in the non-reflection point cloud and whose Euclidean distance is not greater than a set distance threshold is searched and selected as the neighboring sampling point corresponding to the real image sampling point.

10. The lidar calibration method as described in claim 1, characterized in that, Constructing an objective function for the deviation between the real image sampling point and the corresponding neighboring sampling point, including: The objective function is constructed based on the distance deviation between the real image sampling point and the corresponding neighboring sampling point. ;in, This represents the parameters to be optimized in the initial reflective surface equation. This represents the total number of real image sampling points. Indicates in the parameter Under the conditions One real image sampling point; Indicates the first The neighboring sampling points corresponding to each real image sampling point.

Images