Coupling compensation method based on laser measurement

Abstract

The invention discloses a coupling compensation method based on laser measurement, and the method comprises the steps: A, controlling a line laser transmitter to transmit a laser line to a calibration plate, and controlling the calibration plate to move relative to a line laser module; b, under the condition that the calibration plate moves at least two reference positions relative to the line laser module, when the calibration plate moves to one reference position every time, the camera is controlled to collect calibration plane point cloud from the calibration plate, curve fitting and axial symmetry overturning are conducted through the calibration plane point cloud, and a compensation curve equation is obtained; c, when the calibration plate moves to a to-be-calibrated position relative to the line laser module, controlling the camera to collect a to-be-calibrated point cloud from the calibration plate; and calculating a scaling coefficient by using a proportional relation between the point cloud to be calibrated and the point cloud of the calibration plane, and compensating a measured value to be calibrated of the point cloud to be calibrated based on a position relation between the point cloud to be calibrated and the reference position in combination with the scaling coefficient and the compensation curve equation to obtain a final measured value of the point cloud to be calibrated.

Description

Technical Field

[0001] This application relates to the field of sensor technology, and in particular to a coupling compensation method based on laser measurement. Background Technology

[0002] In the process of acquiring images of the target surface of a laser target using a camera in a line laser measurement system, in order to obtain a larger detection range, the camera usually uses a fisheye lens with a larger field of view. Chinese patent CN116503478A discloses a scheme to determine whether distortion has occurred by calculating the ranging error generated by each side of a rectangular label. This makes it easy to detect the characteristic of fisheye lens cameras: the distortion is larger near the edge area, which reduces the camera's recognition accuracy. If the accuracy is improved by reducing the field of view of the fisheye lens, the detection range will be reduced, which will not meet the measurement accuracy requirements of target surface images under different measurement distances and heights, as well as target surface images with object interference. Summary of the Invention

[0003] This application aims to provide a coupling compensation method based on laser measurement, and the specific technical solution is as follows:

[0004] A coupling compensation method based on laser measurement is applied to a line laser module, which includes a line laser emitter and a camera. The coupling compensation method includes: Step A, controlling the line laser emitter to emit a laser line towards a calibration plate and controlling the calibration plate to move relative to the line laser module; then proceeding to Step B; Step B, when the calibration plate moves at least two reference positions relative to the line laser module, whenever the calibration plate moves to a reference position, controlling the camera to acquire a calibration plane point cloud from the calibration plate; then using the calibration plane point cloud to perform curve fitting and axisymmetric flipping to obtain a compensation curve equation; then proceeding to Step C; Step C, whenever the calibration plate moves relative to the line laser module to a position to be calibrated, controlling the camera to acquire a point cloud to be calibrated from the calibration plate; using the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud to calculate a scaling factor, and then based on the positional relationship between the point cloud to be calibrated and the reference position, combined with the scaling factor and the compensation curve equation, compensating the measurement value of the point cloud to be calibrated to obtain the final measurement value of the point cloud to be calibrated.

[0005] Combining steps A to C above, this application utilizes the calibration plane point cloud at multiple reference positions where the calibration plate has moved to obtain multiple compensation curve equations and multiple scaling factors required for error compensation of the point cloud to be calibrated. Then, it combines multiple compensation curve equations and multiple scaling factors to compensate the measurement value of the point cloud to be calibrated. This is adaptable to different height positions or different horizontal distance positions. It adaptively compensates from multiple curve equations or combines curve equations obtained from different reference positions for coupled compensation to obtain the final measurement value of the point cloud to be calibrated. This overcomes the different edge distortion errors caused by the calibration plate at different positions to the camera lens. It can compensate and correct for distortion errors, laser lamp errors, or errors caused by module deformation in different measurement range areas, so that the compensated point cloud improves the edge measurement accuracy of the calibration plate and improves the measurement accuracy and stability of the line laser module.

[0006] Moreover, the compensation curve equations obtained at different reference positions are coupled with the corresponding scaling coefficients for compensation, which has higher accuracy than the compensation curve fitted at a single position and is more robust to measurement in different environments. Thus, the line laser module that performs the coupling compensation method can not only effectively compensate for edge measurement errors when following the movement of the moving carrier, but also adapt to the measurement accuracy at different heights or different detection distances, thus expanding the error compensation application range of the line laser module. Attached Figure Description

[0007] Figure 1 This is a schematic flowchart of a coupling compensation method based on laser measurement, as disclosed in one embodiment of this application. Detailed Implementation

[0008] The technical solutions of the embodiments of this application will be described in detail below with reference to the accompanying drawings. To further illustrate the embodiments, this application provides accompanying drawings. These drawings are part of the disclosure of this invention, mainly used to illustrate the embodiments, and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of this application.

[0009] The calibration of line laser sensors and their use in precision ranging involve camera lenses, such as fisheye lenses used to achieve a wider detection range. Fisheye lenses are characterized by increased distortion and reduced light intake near the edges. Measurement errors are typically suppressed by reducing the field of view, but this reduces measurement flexibility due to the smaller detection range. For example, when a camera acquires images of a fixed calibration plate, including images of the laser target (the target surface of the calibration plate), the accuracy obtained by fitting a small number of point clouds at the same height or detection distance is sufficient. However, when measuring targets at different horizontal distances, heights, and in environments with obstructions, the required accuracy of the acquired target surface images is higher. Multiple states of laser point cloud data are needed to improve robustness and accommodate measurement requirements at different horizontal distances, heights, and in environments with obstructions. Otherwise, when the camera acquires images of the target object in real time and performs real-time ranging, the target object will exhibit different measurement errors relative to the same camera in different environments, affecting measurement accuracy and stability.

[0010] To address the aforementioned technical problems, this application discloses a coupling compensation method based on laser measurement. This method is applied to a line laser module, which has a built-in controller. The controller executes the coupling compensation method to perform coupling compensation during laser measurement using a movable target. Because the target is the object to be measured and is movable, the line laser module can acquire point cloud data of the target at different distances and heights. This point cloud data at different distances and heights is then coupled and processed to compensate for errors, ensuring environmental robustness. The line laser module includes a line laser emitter and a camera. The line laser emitter and camera can combine point cloud data collected from target objects such as planar plates at different locations to perform coupling compensation for camera lens edge distortion errors. In practical applications, the line laser module after implementing the coupling compensation method can be installed on the body of a mobile robot that requires navigation or object ranging.

[0011] See Figure 1 It can be seen that the coupling compensation method includes:

[0012] Step A: Control the line laser emitter to emit laser lines towards the calibration board and control the calibration board to move relative to the line laser module, then execute Step B. The calibration board can move to multiple positions in one direction, or move to multiple positions after changing directions. During the movement of the calibration board relative to the line laser module, the line laser emitter emits laser lines towards the calibration board in real time, and when the calibration board moves relative to the camera to the reference position, the camera is controlled to acquire laser images from within the calibration board in real time, so as to acquire different batches of laser point clouds from the calibration board at different reference positions. A batch of laser point clouds can represent the three-dimensional coordinate information of the location of the calibration board relative to the camera. The reference position is a relative position set in advance according to the calibration accuracy (the calibration position set relative to the camera in one direction).

[0013] The camera captures the laser image formed by the reflection of the laser line through the calibration plate and obtains a batch of laser point clouds through mapping and transformation. This is considered as the point cloud information acquired by the line laser module from the calibration plate, which can be understood as the current planar point cloud obtained from the projected surface of the calibration plate. In specific implementation, each laser point in the laser point cloud is obtained by extracting the brightness center of the laser line from the laser image. For example, a laser line emitted by the line laser emitter is projected onto the calibration plate to form multiple parallel laser stripes. The center point of each parallel laser stripe is the brightness center of the column, which is used as a laser point in the laser point cloud. The laser image can be used to represent the image of the entire calibration plate, covering all edges of the calibration plate. However, it also introduces error factors such as camera edge distortion and laser light intensity, which leads to a decrease in laser image quality. In particular, it cannot adapt to the measurement accuracy at different distances or heights. It is necessary to control the calibration plate to move in the direction that needs to be calibrated, obtain multiple frames of laser images in the direction that needs to be calibrated, and obtain a batch of laser point clouds in each frame of laser image. Subsequently, the batches of laser point clouds are coupled and processed to compensate for the aforementioned error factors.

[0014] Step B: When the calibration plate moves at least two reference positions relative to the line laser module, whenever the calibration plate moves to a reference position, control the camera to collect the calibration plane point cloud from the calibration plate as a calibration plane point cloud at a reference position; then use the calibration plane point cloud to perform curve fitting and axisymmetric flipping to obtain the compensation curve equation; then execute step C.

[0015] If the calibration plate moves to a position at least different from the position of the line laser module, and is also within the camera's field of view (capable of focusing and imaging the laser point cloud formed by reflection from the calibration plate surface), then the position the calibration plate has moved to is taken as a reference position. The point cloud acquired by the camera from the calibration plate at the reference position, which is distributed in a planar manner, can be considered a calibration planar point cloud. The reference position closest to the camera in the horizontal or vertical direction can be the first reference position the calibration plate has moved to, and the reference position farthest from the camera in the horizontal or vertical direction can be the last reference position the calibration plate has moved to. The interval between two adjacent reference positions the calibration plate has moved to can be fixed. If there are obstacles interfering with the calibration plate and the camera, causing the point cloud acquired by the camera in real time to not be distributed in a planar manner, then the point cloud that is not distributed in a planar manner cannot be considered a calibration planar point cloud.

[0016] To improve calibration / compensation accuracy, step B performs curve fitting on the calibration plane point cloud acquired at each reference position, forming multiple fitting equations to characterize the bending trend of the point cloud at different distances. Then, each fitting equation is axisymmetrically flipped to obtain a corresponding compensation curve equation, thus forming multiple compensation curve equations. This provides a reverse compensation amount for subsequent coupling compensation of laser points corresponding to different measurements to be calibrated. In some embodiments, the calibration plane point cloud can be a pre-calibrated point cloud, which can be understood as the point cloud acquired when the calibration plate is moved to a reference position and then error-compensated, and is considered to be distributed within the projection plane of the calibration plate.

[0017] Step C: Whenever the calibration board moves relative to the line laser module to a position to be calibrated, the camera is controlled to acquire the point cloud to be calibrated from the calibration board; the scaling factor is calculated using the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud; then, based on the positional relationship between the point cloud to be calibrated and the reference position, the scaling factor and the compensation curve equation are combined to compensate the measurement value to be calibrated of the point cloud to be calibrated, and the final measurement value of the point cloud to be calibrated is obtained, including the final measurement value of each laser point of the point cloud to be calibrated.

[0018] After determining the positional relationship between the point cloud to be calibrated acquired at the calibration location and the at least two reference locations mentioned in step B in step C, the calibration measurement values ​​of the point cloud to be calibrated are compensated by combining the scaling factor and the compensation curve equation. Specifically, the calibration measurement values ​​of the point cloud to be calibrated are compensated point by point to obtain the final measurement value of each laser point in the point cloud to be calibrated. The calibration measurement values ​​of the point cloud to be calibrated are specifically the calibration measurement values ​​of each laser point in the point cloud to be calibrated. The calibration measurement values ​​of the point cloud to be calibrated include height information and distance information, which can be represented by the coordinate values ​​of the vertical coordinate axis and the horizontal coordinate axis of the spatial coordinate system, respectively.

[0019] The positional relationship between the point cloud to be calibrated and the reference position described in step B is not necessarily equivalent to the positional relationship between the position to be calibrated and the calibration plane point cloud described in step B due to lens distortion, and error compensation is required. It should be noted that the point cloud to be calibrated acquired in step C can be considered as a point cloud distributed on a curved surface, or as a non-planar point cloud generated by interference from obstacles between the camera and the calibration plate; both are sets of laser points that need to be calibrated.

[0020] The laser point at the calibration location or in the calibration point cloud can be located between two adjacent reference locations; or, the distance between the laser point at the calibration location or in the calibration point cloud and the camera can be less than the distance between any reference location and the camera; or, the height of the laser point at the calibration location or in the calibration point cloud and the camera can be less than the height of any reference location and the camera; or, the distance between the laser point at the calibration location or in the calibration point cloud and the camera can be greater than the distance between any reference location and the camera; or, the height of the laser point at the calibration location or in the calibration point cloud and the camera can be greater than the height of any reference location and the camera.

[0021] To ensure compatibility with measurement accuracy at different distances or heights, step C can first determine the relationship between the point cloud to be calibrated acquired at the calibration location and the measurement value to be calibrated at the reference location. Then, the scaling factor is introduced into the compensation curve equation obtained at the reference location to directly compensate the measurement value to be calibrated in the point cloud to obtain the final measurement value of the point cloud to be calibrated. Alternatively, a weighted average can be calculated by combining the compensation curve equations corresponding to the calibration plane point clouds acquired at two adjacent reference locations (which contribute to the compensation amount required for the measurement value to be calibrated at a single reference location) and the mean of the measurement values ​​to be calibrated in the same calibration plane point cloud (which contributes to the weight) to obtain a weighted compensation result, which serves as the final measurement value of each laser point in the point cloud to be calibrated.

[0022] Therefore, after controlling the calibration plate to move through all the required reference positions in step B, at least two reference positions adjacent to the laser point of the point cloud to be calibrated can be obtained before proceeding to step C. In step C, the scaling factor and compensation curve equation to be calculated are determined based on the relationship between the measurement value to be calibrated of the point cloud and the preset maximum or minimum measurement value to be calibrated. This determines the specific compensation method for the measurement value to be calibrated of the point cloud, thereby affecting the measurement accuracy of the point cloud to be calibrated. After determining the reference position to be compensated, the collected calibration plane point cloud, and the compensation curve equation to be obtained based on step B, step C uses a pre-calculated scaling factor (here, the scaling factor is used to represent the ratio between the pose information of the point to be calibrated and the pose information of the laser point in the calibration board (considered as the pre-calibrated center point), and there are scaling factors in different dimensions) to introduce the compensation curve equation obtained by the calibration board at the same reference position. Then, compensation processing can be performed on each original laser point in the point cloud to be calibrated, including substituting a coordinate value of each original laser point into the independent variable of the compensation curve equation after introducing the scaling factor. The obtained function result can directly offset or offset the measurement error of the original laser point to be calibrated by performing a weighted average calculation after coupling the measurement values ​​to be calibrated.

[0023] Combining steps A to C above, this application utilizes the calibration plane point cloud at multiple reference positions where the calibration plate has moved to obtain multiple compensation curve equations and multiple scaling factors required for error compensation of the point cloud to be calibrated. Then, it combines multiple compensation curve equations and multiple scaling factors to compensate the measurement value of the point cloud to be calibrated. This is adaptable to different height positions or different horizontal distance positions. It adaptively compensates from multiple curve equations or combines curve equations obtained from different reference positions for coupled compensation to obtain the final measurement value of the point cloud to be calibrated. This overcomes the different edge distortion errors caused by the calibration plate at different positions to the camera lens. It can compensate and correct for distortion errors, laser lamp errors, or errors caused by module deformation in different measurement range areas, so that the compensated point cloud improves the edge measurement accuracy of the calibration plate and improves the measurement accuracy and stability of the line laser module.

[0024] Moreover, the compensation curve equations obtained at different reference positions are coupled with the corresponding scaling coefficients for compensation, which has higher accuracy than the compensation curve fitted at a single position and is more robust to measurement in different environments. Thus, the line laser module that performs the coupling compensation method can not only effectively compensate for edge measurement errors when following the movement of the moving carrier, but also adapt to the measurement accuracy at different heights or different detection distances, thus expanding the error compensation application range of the line laser module.

[0025] As one embodiment, in step C, the method for compensating the measurement value of the point cloud to be calibrated based on the positional relationship between the point cloud to be calibrated and the reference position, combined with the scaling factor and the compensation curve equation, to obtain the final measurement value of the point cloud to be calibrated includes:

[0026] The positional relationship between a laser point in the point cloud to be calibrated and a target calibration interval is determined. The target calibration interval is the region between a first and a second measurement value to be calibrated, with the two ends of the target calibration interval being a first reference position and a second reference position, respectively; the first measurement value to be calibrated is less than the second measurement value to be calibrated. In practice, the positional relationship between each laser point in the point cloud to be calibrated and the target calibration interval can be determined by comparing the measurement value to be calibrated with the first and second measurement values.

[0027] It should be noted that, within the camera coordinate system, the measurement value to be calibrated at the first reference position is the first measurement value to be calibrated. This measurement value can be the horizontal distance or height difference between the first reference position and the camera, and can be represented using coordinates on the corresponding coordinate axes. Similarly, the measurement value to be calibrated at the second reference position is the second measurement value to be calibrated. This measurement value can be the horizontal distance or height difference between the second reference position and the camera, and can also be represented using coordinates on the corresponding coordinate axes.

[0028] The reference positions that the calibration board needs to move to are all located between the first reference position and the second reference position, that is, all the reference positions that the calibration board needs to move to are within the target calibration range. The reference positions that the calibration board needs to move to include the first reference position and the second reference position, both of which are the calibration positions required to calibrate the point cloud acquired by the camera. Within the target calibration range, if there are no other reference positions that the calibration board needs to move to besides the first and second reference positions, the first and second reference positions form two adjacent reference positions.

[0029] In some embodiments, from a horizontal distance perspective, the first reference position and the second reference position are respectively the reference position closest to the camera at horizontal distance and the reference position furthest from the camera at horizontal distance. The first reference position can be understood as the end of the target calibration interval with the smallest measurement value to be calibrated, and the second reference position can be understood as the end of the target calibration interval with the largest measurement value to be calibrated. Schematic, with the camera (camera coordinate system) as the origin, the measurement value to be calibrated is the coordinate of the z-axis. From a height (vertical direction) perspective, the first reference position and the second reference position are respectively the reference position with the smallest height relative to the camera and the reference position with the largest height relative to the camera.

[0030] It is worth noting that before performing steps B and C, the calibration measurement values ​​at both the first and second reference positions are known. The calibration measurement values ​​for each laser point in the calibration point cloud and at the calibration positions can be obtained in real time.

[0031] If it is determined that a laser point in the point cloud to be calibrated is outside the target calibration interval, the end with the smallest difference between the laser point and the laser point to be calibrated measurement value in the target calibration interval is selected as the target reference position. In this embodiment, it can be determined whether the same laser point is outside the target calibration interval by judging the relationship between the laser point to be calibrated measurement value and the first and second laser point to be calibrated measurement values.

[0032] If it is determined that the measured value to be calibrated of a laser point in the point cloud to be calibrated is less than the first measured value to be calibrated, then it is determined that the laser point in the point cloud to be calibrated is located outside the target calibration interval, and the end of the target calibration interval with the smallest difference between the measured value to be calibrated and the laser point in the point cloud to be calibrated is the first reference position, that is, the target reference position is the first reference position; then, the measured value to be calibrated of the laser point in the point cloud to be calibrated is compensated by combining the scaling factor of the target reference position and the compensation curve equation at the target reference position, to obtain the final measured value of the laser point in the point cloud to be calibrated. During the execution of step C, before determining that the target reference position is the first reference position, the first horizontal scaling factor, the first vertical scaling factor, and the first amplitude scaling factor can be calculated based on the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud at the first reference position, respectively. The scaling factor of the first reference position includes the first horizontal scaling factor, the first vertical scaling factor, and the first amplitude scaling factor, and the corresponding scaling factor of the target reference position includes the horizontal scaling factor, the vertical scaling factor, and the amplitude scaling factor. Therefore, the method of compensating the measurement value of a laser point to be calibrated by combining the scaling factor of the target reference position and the compensation curve equation at the target reference position is equivalent to the method of compensating the measurement value of a laser point in the point cloud to be calibrated by combining the scaling factor of the first reference position and the compensation curve equation at the first reference position. Specifically, it includes substituting the product of the coordinates of the laser point in the width direction of the point cloud to be calibrated and the lateral scaling factor of the target reference position into the compensation curve equation at the first reference position to obtain a pre-compensation result, wherein the coordinates of the laser point in the width direction of the point cloud to be calibrated are the product of the scaling factor of the target reference position and the compensation curve equation at the first reference position. The coordinates of the calibration point cloud in the width direction are determined. Then, the product of the pre-compensation result, the first longitudinal scaling factor, and the first amplitude scaling factor is added to the calibration measurement value of a laser point in the point cloud to be calibrated. This performs a compensation operation on the calibration measurement value of the laser point in the point cloud to be calibrated, obtaining the final measurement value of the laser point in the point cloud to be calibrated. This compensation operation is repeated for each laser point in the point cloud to be calibrated, thereby compensating the calibration measurement value of the point cloud point by point. This completes the compensation processing of the center points of each column of laser stripes in the laser image, obtaining the compensated coordinates of each laser point in the laser point cloud.

[0033] Schematic, let the first reference position be a horizontal position relative to the camera at a distance from the first measurement value to be calibrated. Let X be the coordinate of a laser point in the point cloud to be calibrated in the width direction, representing the coordinate on the horizontal axis of the spatial coordinate system. Let Z be the measurement value to be calibrated for a laser point in the point cloud to be calibrated. 原始The coordinates on the vertical axis of the spatial coordinate system are used to represent the coordinates of the measurements to be compensated / calibrated. The compensation curve equation at the first reference position is Z = f1(x), which serves as the compensation function required for calibration. The first lateral scaling factor, the first longitudinal scaling factor, and the first amplitude scaling factor can be calculated based on the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud at the first reference position. 原始 If the value is less than the first measurement to be calibrated (which is the minimum measurement to be calibrated), the final measurement value Z of the laser point is... 补偿 =Z 原始 +f 1 (X * first horizontal scaling factor) * first vertical scaling factor * first amplitude scaling factor. Therefore, the aforementioned embodiment utilizes the calibration plane point cloud acquired at the reference position, the scaling factor of the reference position, and the compensation curve equation corresponding to the reference position to perform point-by-point compensation on the point cloud to be calibrated. This achieves compensation to obtain the final measurement value of the point cloud to be calibrated when the measured value to be calibrated is less than the first measured value to be calibrated, thereby improving measurement accuracy when the point cloud to be calibrated is close to the camera.

[0034] If it is determined that the measured value to be calibrated of a laser point in the point cloud to be calibrated is greater than the second measured value to be calibrated, then it is determined that the laser point in the point cloud to be calibrated is located outside the target calibration interval, and the end of the target calibration interval with the smallest difference between the measured value to be calibrated and the laser point in the point cloud to be calibrated is the second reference position, and the target reference position is determined to be the second reference position. Then, the measured value to be calibrated of a laser point in the point cloud to be calibrated is compensated by combining the scaling factor of the target reference position and the compensation curve equation at the target reference position, to obtain the final measured value of the laser point in the point cloud to be calibrated. During the execution of step C, before determining that the target reference position is the second reference position, the second lateral scaling factor, the second longitudinal scaling factor, and the second amplitude scaling factor can be calculated based on the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud at the second reference position, respectively. The scaling factor of the second reference position includes the second lateral scaling factor, the second longitudinal scaling factor, and the second amplitude scaling factor, and the corresponding scaling factor of the target reference position includes the lateral scaling factor, the longitudinal scaling factor, and the amplitude scaling factor. Therefore, the method of compensating the measurement value of a laser point to be calibrated by combining the scaling factor of the target reference position and the compensation curve equation at the target reference position is equivalent to the method of compensating the measurement value of a laser point in the point cloud to be calibrated by combining the scaling factor of the second reference position and the compensation curve equation at the second reference position. Specifically, it includes substituting the product of the coordinates of the laser point in the width direction of the point cloud to be calibrated and the lateral scaling factor of the target reference position into the compensation curve equation at the second reference position to obtain a pre-compensation result, wherein the coordinates of the laser point in the width direction of the point cloud to be calibrated are the product of the scaling factor of the target reference position and the compensation curve equation at the second reference position. The coordinates of the calibration point cloud in the width direction are determined. Then, the product of the pre-compensation result, the second longitudinal scaling factor, and the second amplitude scaling factor is added to the calibration measurement value of a laser point in the point cloud to be calibrated. This performs a compensation operation on the calibration measurement value of the laser point in the point cloud to be calibrated, obtaining the final measurement value of the laser point in the point cloud to be calibrated. This compensation operation is repeated for each laser point in the point cloud to be calibrated, thereby compensating the calibration measurement value of the point cloud to be calibrated point by point. This completes the compensation processing of the center points of each column of laser stripes in the laser image, obtaining the compensated coordinates of each laser point in the laser point cloud.

[0035] Schematic, let the second reference position be a reference position in the horizontal direction that is a distance from the camera from the second measurement value to be calibrated. Let X be the coordinate of a laser point in the point cloud to be calibrated in the width direction, representing the coordinate on the horizontal axis of the spatial coordinate system. Let Z be the measurement value to be calibrated for a laser point in the point cloud to be calibrated. 原始The coordinates on the vertical axis of the spatial coordinate system are also the coordinate values ​​used for the measurements to be compensated / calibrated; the compensation curve equation at the second reference position is Z = f2(x), which serves as the compensation function required for calibration. Then, based on the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud at the second reference position, the second lateral scaling factor, the second longitudinal scaling factor, and the second amplitude scaling factor can be calculated respectively; Z 原始 When the value is greater than the second measurement to be calibrated (which is the largest measurement to be calibrated), the final measurement value Z of the laser point is... 补偿 =Z 原始 +f2(X*second horizontal scaling factor)*second vertical scaling factor*second amplitude scaling factor. Therefore, the aforementioned embodiment utilizes the calibration plane point cloud collected at the reference position, the scaling factor of the reference position, and the compensation curve equation corresponding to the reference position to perform point-by-point compensation on the point cloud to be calibrated, so as to obtain the final measurement value of the point cloud to be calibrated when the measurement value to be calibrated of the point cloud to be calibrated is greater than the second measurement value to be calibrated.

[0036] It should be noted that the product of the coordinates of a laser point in the width direction and the first horizontal scaling factor in the point cloud to be calibrated represents scaling the coordinate value in the width direction. Furthermore, the product of the pre-compensation result, the vertical scaling factor, and the amplitude scaling factor represents scaling in the coordinate axis direction corresponding to the measurement value to be calibrated, respectively, from the dimensions of extreme coordinate values ​​and the average coordinate value. This achieves the processing of the compensation curve equation at the target reference position using scaling factors in three directions of the target reference position. Thus, the compensation amount required for the measurement value to be calibrated of each laser point in the point cloud to be calibrated is obtained by using the calibration plane point cloud at the reference position closest to the point cloud to be calibrated. Finally, the compensation amount required for the measurement value to be calibrated of each laser point in the point cloud to be calibrated is sequentially controlled to participate in the additive compensation to obtain the final measurement value of each laser point in the point cloud to be calibrated.

[0037] If a laser point in the point cloud to be calibrated is determined to be within the target calibration range, then the scaling factor of the first reference position, the scaling factor of the second reference position, the compensation curve equation at the first reference position, and the compensation curve equation at the second reference position are combined to perform compensation calculation on the measurement value to be calibrated for the laser point, thereby obtaining the pre-calibration result of the laser point at the first reference position and the pre-calibration result of the laser point at the second reference position. Then, using the mean of the measurement values ​​to be calibrated in the calibration plane point cloud at the first reference position and the mean of the measurement values ​​to be calibrated in the calibration plane point cloud at the second reference position, a weighted average is calculated on the pre-calibration result of the laser point at the first reference position and the pre-calibration result of the laser point at the second reference position to obtain the final measurement value of the laser point.

[0038] Specifically, if the measured value of a laser point in the point cloud to be calibrated is greater than or equal to the first measured value to be calibrated, and the measured value of a laser point in the point cloud to be calibrated is less than or equal to the second measured value to be calibrated, then the laser point in the point cloud to be calibrated is determined to be within the target calibration interval. It should be noted that within the target calibration interval, two reference positions are arranged in order of distance from the camera from closest to farthest along a target coordinate axis. The reference position closest to the camera is the first reference position, and the reference position farthest from the camera is the second reference position. The target coordinate axis direction is the direction of the coordinate axis where the measured value to be calibrated is located, indicating the direction of movement of the calibration plate. The scaling factor for each reference position includes a lateral scaling factor, a longitudinal scaling factor, and an amplitude scaling factor. The scaling factor for the first reference position includes a first lateral scaling factor, a first longitudinal scaling factor, and a first amplitude scaling factor; the scaling factor for the second reference position includes a second lateral scaling factor, a second longitudinal scaling factor, and a second amplitude scaling factor.

[0039] When it is determined that a laser point in the point cloud to be calibrated is within the target calibration range, the method of calculating the compensation of the laser point's measurement value by combining the scaling factor of the first reference position, the scaling factor of the second reference position, the compensation curve equation at the first reference position, and the compensation curve equation at the second reference position, to obtain the pre-calibration result of the laser point at the first reference position and the pre-calibration result of the laser point at the second reference position, includes:

[0040] Substitute the product of the coordinates of a laser point in the width direction of the point cloud to be calibrated and the first lateral scaling factor into the compensation curve equation at the first reference position to obtain the first pre-compensation result; then add the product of the first pre-compensation result, the first longitudinal scaling factor, and the first amplitude scaling factor to the calibration measurement value of a laser point in the point cloud to be calibrated to obtain the laser point pre-calibration result at the first reference position; wherein, the lateral scaling factor, the longitudinal scaling factor, and the amplitude scaling factor at the first reference position are denoted as the first lateral scaling factor, the first longitudinal scaling factor, and the first amplitude scaling factor, respectively.

[0041] Substitute the product of the coordinates of a laser point in the width direction of the point cloud to be calibrated and the second lateral scaling factor into the compensation curve equation at the second reference position to obtain the second pre-compensation result; then add the product of the second pre-compensation result, the second longitudinal scaling factor, and the second amplitude scaling factor to the calibration measurement value of a laser point in the point cloud to be calibrated to obtain the laser point pre-calibration result at the second reference position; wherein, the lateral scaling factor, longitudinal scaling factor, and amplitude scaling factor at the second reference position are denoted as the second lateral scaling factor, the second longitudinal scaling factor, and the second amplitude scaling factor, respectively.

[0042] For each laser point in the point cloud to be calibrated, the pre-calibration results for the laser point at the first reference position and the pre-calibration results for the laser point at the second reference position need to be calculated. This allows for point-by-point calculation to obtain the pre-calibration results for each reference position, i.e., to obtain the pre-calibration results for each reference position within the target calibration interval.

[0043] During step C, if it is determined that a laser point in the point cloud to be calibrated is located within the target calibration interval, the method for calculating the weighted average of the pre-calibration results of the laser point at the first reference position and the pre-calibration results of the laser point at the second reference position using the average of the measured values ​​to be calibrated in the calibration plane point cloud at the first reference position and the average of the measured values ​​to be calibrated in the calibration plane point cloud at the second reference position includes: recording the average of the measured values ​​to be calibrated in the calibration plane point cloud at the first reference position as the first reference amplitude average, that is, the average of the measured values ​​to be calibrated of each laser point in the calibration plane point cloud at the first reference position is equal to the first reference amplitude average; and recording the absolute value of the difference between the measured value to be calibrated of a laser point in the point cloud to be calibrated and the first reference amplitude average as the first relative compensation difference. Furthermore, the mean value of the measurement values ​​to be calibrated of the calibration plane point cloud at the second reference position is recorded as the second reference amplitude mean value, that is, the average value of the measurement values ​​to be calibrated of each laser point in the calibration plane point cloud at the second reference position is equal to the second reference amplitude mean value; and the absolute value of the difference between the measurement value to be calibrated of a laser point in the point cloud to be calibrated and the second reference amplitude mean value is recorded as the second relative compensation difference value. Then, the absolute value of the difference between the average value of the first reference amplitude and the average value of the second reference amplitude is recorded as the interval compensation difference; the ratio between the first relative compensation difference and the interval compensation difference is recorded as the second weight, and the ratio between the second relative compensation difference and the interval compensation difference is recorded as the first weight; then, the laser point pre-calibration result at the first reference position is multiplied by the second weight to obtain the first weighted value; the laser point pre-calibration result at the second reference position is multiplied by the first weight to obtain the second weighted value; then, the first weighted value and the second weighted value are added to obtain the final measurement value of a laser point in the point cloud to be calibrated. This realizes the weighted average calculation of the laser point pre-calibration result at the first reference position and the laser point pre-calibration result at the second reference position. For the measurement value to be calibrated of each laser point in the point cloud to be calibrated, the weighted average calculation is required using the laser point pre-calibration result at the first reference position and the laser point pre-calibration result at the second reference position. This realizes the pre-calibration result of each reference position is obtained by calculating point by point, that is, the pre-calibration result of each reference position within the target calibration interval is obtained. This allows for weighted averaging of the mean values ​​of the point clouds on different calibration planes, balancing the influence of extreme values ​​that may exist in the pre-calibration results of laser points at different reference positions. This overcomes the different edge distortion errors caused by the calibration plate at different distances to the camera lens, and can compensate for and correct errors caused by distortion, laser lamp, or module deformation in different measurement ranges. The compensated point cloud improves the edge measurement accuracy of the calibration plate and enhances the measurement accuracy and stability of the line laser module.

[0044] Schematic, N equals 2, and i takes values ​​of 1 and 2 respectively; let the mean of the measurements to be calibrated in the calibration plane point cloud at the first reference position be the average distance of the calibration plane point cloud at the first reference position, denoted as Z. 距离1均值 Let the mean of the measurements to be calibrated in the calibration plane point cloud at the second reference position be the average distance of the calibration plane point cloud at the second reference position, denoted as Z. 距离2均值 If a laser point in the point cloud to be calibrated is located within the target calibration interval, the calibration measurement value of each laser point in the point cloud to be calibrated can be used using Z. 原始 Z indicates 原始 Greater than Z 距离1均值 And Z 原始 Less than Z 距离2均值 The product of the coordinate X of a laser point in the width direction and the first lateral scaling factor in the point cloud to be calibrated is substituted into the compensation curve equation at the first reference position to obtain the first pre-compensation result, i.e., f1(X*first lateral scaling factor); then, the pre-calibration result of the laser point at the first reference position is Z. 补偿1 =Z 原始 +f1(X*first horizontal scaling factor)*first vertical scaling factor*first amplitude scaling factor. Similarly, the product of the coordinate X of a laser point in the width direction and the second horizontal scaling factor in the point cloud to be calibrated is substituted into the compensation curve equation at the second reference position to obtain the second pre-compensation result, i.e., f2(X*second horizontal scaling factor); then, the pre-calibration result of the laser point at the second reference position is Z. 补偿2 =Z 原始 +f2(X*second horizontal scaling factor)*second vertical scaling factor*second amplitude scaling factor. Set the first relative compensation difference to Z. 原始 -Z 距离1均值 Set the second relative compensation difference to Z. 距离2均值 -Z 原始 Set the interval compensation difference to Z. 距离2均值 -Z 距离1均值 Then set the first weight to Set the second weight to Finally, the final measured value of a laser point in the point cloud to be calibrated is calculated to be Z. 补偿 =Z 补偿1 *Z 权重2 +Z 补偿2 *Z 权重1 Therefore, a weighted average is calculated by combining the compensation curve equations (which contribute to the compensation amount required for the measurement value to be calibrated at a single reference position) corresponding to the calibration plane point clouds collected at the first and second reference positions, and the mean of the measurement values ​​to be calibrated in the same calibration plane point cloud (which contributes to the weight), to obtain the weighted compensation result, which serves as the final measurement value of each laser point in the point cloud to be calibrated.

[0045] As one embodiment, the method for calculating the scaling factor based on the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud includes, assuming that all laser points within the point cloud to be calibrated are distributed on the same plane, and it is determined that the point cloud acquired by the line laser module from the calibration board is not interfered with by obstacles, and the scaling factors generated by the point cloud to be calibrated relative to the calibration plane point cloud are all planar scaling factors, then the scaling factors in different dimensions are calculated separately, and the following exists:

[0046] Calculate the coordinate range of the point cloud to be calibrated in the width direction to obtain the current lateral range. In the spatial coordinate system of the camera, the width direction of the point cloud to be calibrated is perpendicular to the optical axis of the camera and parallel to the imaging plane of the camera. The coordinates in the width direction of each laser point in the point cloud to be calibrated can be recorded as the abscissa. The coordinate range of the point cloud to be calibrated in the width direction is the difference between the maximum and minimum abscissas among the abscissas of each laser point in the point cloud to be calibrated.

[0047] Calculate the range of the measurement values ​​to be calibrated in the point cloud to obtain the current longitudinal range. Whether in the spatial coordinate system or projected onto a coordinate plane, each laser point in the point cloud to be calibrated has a measurement value to be calibrated. This value can be the horizontal distance or height of each laser point in the point cloud to be calibrated relative to the optical center of the camera (the origin of the coordinate system), and can be represented by coordinate values. The range of the measurement values ​​to be calibrated in the point cloud to be calibrated is the difference between the largest and smallest measurement values ​​among all the laser points in the point cloud to be calibrated.

[0048] The mean of the calibration measurements of the point cloud to be calibrated is calculated to obtain the current mean amplitude. Each laser point in the point cloud to be calibrated has a calibration measurement value, which can be the horizontal distance or height of each laser point in the point cloud to be calibrated relative to the optical center (origin of the coordinate system) of the camera. Due to the distortion at the lens edge, even without obstacles, the calibration measurement values ​​of any two laser points in the laser point cloud acquired by the camera from the same calibration plate are not equal. That is, there is a maximum and a minimum calibration measurement value. In addition to calculating the range of the calibration measurements of the point cloud to be calibrated, the mean of the calibration measurements of the point cloud to be calibrated is also calculated to represent the average amplitude of the fluctuation of the calibration measurements of the point cloud to be calibrated, which is also equivalent to the curvature amplitude of the surface on which the point cloud to be calibrated is located. The closer the camera is to the calibration plate, the smaller the mean of the calibration measurements of the point cloud to be calibrated.

[0049] It should be noted that the calibration plane point cloud refers to the coordinate parameters required for calibration at a reference position before compensating the point cloud to be calibrated. The point cloud to be calibrated is not necessarily the calibration plane point cloud.

[0050] During step C, it may also be necessary to determine whether there are obstacles between the line laser module and the calibration board. The determination method includes: whenever the calibration board is located at a calibration position, it is determined whether each laser point in the calibration point cloud acquired by the camera in real time is coplanar; if so, it is determined that each laser point in the calibration point cloud is distributed on the same plane, that is, the camera directly acquires the calibration point cloud from the surface of the calibration board; otherwise, it is determined that each laser point in the calibration point cloud is not distributed on the same plane, and the camera acquires the calibration point cloud from the surface of the obstacle in front of the calibration board.

[0051] For a reference position, the coordinate range of the calibration plane point cloud in the width direction is calculated to obtain the reference lateral range. The coordinate range of the calibration plane point cloud in the width direction is the difference between the maximum and minimum abscissas of each laser point in the calibration plane point cloud. The ratio between the current lateral range and the reference lateral range is then set as the lateral scaling factor for the reference position. This ratio corresponds to a proportional relationship between the point cloud to be calibrated and the calibration plane point cloud, representing the coefficient of variation of the extreme coordinate values ​​of the spatial region where the point cloud to be calibrated is located relative to the calibration plane point cloud in the width direction. It is significantly affected by the two extreme values ​​of the coordinates in the width direction, while ignoring most of the intermediate values.

[0052] It should be noted that the width of the laser stripe in the laser image is the pixel distance formed in the column direction of the laser image after the laser line is reflected back to the line laser module; the width direction of the laser line in the laser image is parallel to each column of the laser image. The width direction of the laser stripe is pre-calibrated to be parallel to each column of the laser image using the intrinsic parameter matrix of the line laser module, and the width direction of the calibration plate can also be the pre-calibrated width direction of the laser stripe; the length direction of the laser stripe is perpendicular to the width direction of the laser stripe. The laser stripe is distributed longitudinally within multiple columns of the laser image, which can be understood as being composed of multiple longitudinal light bars arranged along the length direction of the laser stripe.

[0053] For a reference position, the range of the measurement values ​​to be calibrated in the calibration plane point cloud is calculated to obtain the reference longitudinal range. The range of the measurement values ​​to be calibrated in the calibration plane point cloud is the difference between the largest and smallest measurement values ​​among the laser points in the calibration plane point cloud. The ratio between the current longitudinal range and the reference longitudinal range is then set as the longitudinal scaling factor for the reference position. This ratio corresponds to a proportional relationship between the point cloud to be calibrated and the calibration plane point cloud, representing the extreme coordinate value variation parameter of the spatial region containing the point cloud to be calibrated relative to the calibration plane point cloud in the direction of the measurement value to be calibrated.

[0054] For a reference position, the mean value of the measurement values ​​to be calibrated in the calibration plane point cloud is calculated to obtain the reference amplitude mean value. The mean value of the measurement values ​​to be calibrated in the calibration plane point cloud can describe the overall curvature of the curve formed by each laser point in the calibration plane point cloud relative to the optical center of the camera in a plane. Then, the ratio between the current amplitude mean value and the reference amplitude mean value is set as the amplitude scaling factor of the reference position. The ratio between the current amplitude mean value and the reference amplitude mean value corresponds to a proportional relationship between the point cloud to be calibrated and the calibration plane point cloud, representing the coefficient of change of the average coordinate value of the spatial region where the point cloud to be calibrated is located relative to the spatial region where the calibration plane point cloud is located in the direction of the measurement values ​​to be calibrated.

[0055] Based on the foregoing embodiments, when it is determined that a laser point in the point cloud to be calibrated is outside the target calibration range, the scaling factor at the reference position can be the scaling factor at the target reference position.

[0056] The scaling factor for a reference position includes a horizontal scaling factor, a vertical scaling factor, and an amplitude scaling factor. The scaling factor can represent the reduction or magnification factor of the spatial surface containing the point cloud acquired at the reference position relative to the calibration plate plane. The reference position can be a first reference position or a second reference position.

[0057] As another embodiment, the method for calculating the scaling factor based on the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud further includes, when the point cloud to be calibrated is not distributed on a plane, and it is determined that the point cloud acquired by the line laser module from the calibration plate is affected by obstacles, the scaling factors in different dimensions are set to be equal. The calculation method for the scaling factors in different dimensions is as follows: at a reference position, the mean value of the measurement values ​​to be calibrated in the calibration plane point cloud is calculated. In this case, the calibration plane point cloud is a point cloud distributed on a plane, pre-acquired from a calibration plate at a reference position, and the acquisition is completed under conditions of no obstruction. The mean value of the measurement values ​​to be calibrated in the calibration plane point cloud represents the average amplitude of the fluctuation of the measurement values ​​to be calibrated in the point cloud, which is also equivalent to the curvature amplitude of the surface where the point cloud to be calibrated is located, providing a reference basis for calculating the scaling factor of the point cloud to be calibrated. For each laser point within the point cloud to be calibrated, the ratio between the measured value of the laser point to be calibrated and the mean of the measured values ​​to be calibrated in the calibration plane point cloud is calculated. The calculated ratio is then set as the horizontal scaling factor, vertical scaling factor, or amplitude scaling factor for a reference position, thereby calculating the scaling factor for the reference position point by point. The horizontal scaling factor is equal to the vertical scaling factor, and the vertical scaling factor is equal to the amplitude scaling factor. These scaling factors are all non-planar scaling factors and are the scaling factors at the reference position where the calibration plane point cloud is located.

[0058] It is worth noting that in this embodiment, the mean of the measurement values ​​to be calibrated of the point cloud to be calibrated is not used when calculating the scaling factor point by point, nor can the range be used, because the occlusion effect of obstacles will cause the mean and extreme value information to fail to reflect the curvature of the spatial plane where the point cloud is located caused by the distortion of the camera lens itself, and instead introduce error interference.

[0059] In summary, when calibrating the line laser module (e.g., compensating for errors in the acquired point cloud) using a fixed plane as the calibration board, scaling factors are calculated from different directions. This involves calculating the range and mean, and then performing ratio calculations to observe scaling deformation in different directions, thus improving measurement accuracy. However, when the line laser module is used for autonomous obstacle detection, it only obtains a single scaling factor by calculating the ratio between the measured value to be calibrated and the mean of the measured values ​​to be calibrated on the calibration plane point cloud, without distinguishing between directions.

[0060] As one embodiment, in step B, the method for obtaining the compensation curve equation by performing curve fitting and axisymmetric processing using the calibration plane point cloud includes: performing curve fitting using the calibration plane point cloud collected at the reference position to obtain the curve equation to be calibrated; setting a target axis of symmetry based on the extreme value of the curve equation to be calibrated, and then using the target axis of symmetry to control the curve equation to be calibrated to perform axisymmetric flipping to obtain the compensation curve equation at the reference position; wherein, when the calibration plane point cloud is derived from the calibration plate at the first reference position, the compensation curve equation obtained is at the first reference position; when the calibration plane point cloud is derived from the calibration plate at the second reference position, the compensation curve equation obtained is at the second reference position. The compensation curve equation at the reference position includes the compensation curve equation at the target reference position. Specifically, when it is determined that a laser point in the point cloud to be calibrated is outside the target calibration range, the scaling factor at the reference position can be the scaling factor at the target reference position.

[0061] In step B, the camera's hardware parameters can be set to the standard projection relationship of the camera, and this is established in a distortion-free image coordinate space. The laser image acquired by the camera may be an image with edge distortion. The discrete laser point sequence variation characteristics, laser line arrangement characteristics, or laser stripe distribution characteristics will reflect the bending trend of the point cloud caused by edge distortion errors. Therefore, step B combines the camera's hardware parameters to set a reference plane and performs curve fitting in the reference plane to obtain the equation of the curve to be calibrated to reflect the bending trend of the calibration plane point cloud in the laser image, including errors such as laser stripe deformation and edge position distortion.

[0062] In step B, the maximum or minimum point is obtained by calculating the extreme value of the curve equation to be calibrated. The edge distortion error is concentrated on the laser points distributed at the edge position, which is reflected in the coordinate changes of the maximum or minimum point. Therefore, the target axis of symmetry is set by the extreme value of the curve equation to be calibrated, and then the curve equation to be calibrated is axially symmetrically flipped with the target axis of symmetry as a reference to obtain the compensation curve equation at the reference position. Specifically, the compensation value of the center point of each column of laser stripes can be obtained from the compensation curve equation, including the compensation value in the distance direction between the camera and the calibration plate and the compensation value in the height direction of the calibration plate relative to the ground, forming the compensation information in the three-dimensional point cloud space. In step C, the scaling factor is combined to perform error compensation on the point cloud to be calibrated.

[0063] The coupling compensation method further includes setting a coordinate axis parallel to the camera optical axis as a first coordinate axis to be calibrated, and marking the coordinate values ​​on the first coordinate axis to be calibrated as first coordinates; setting a coordinate axis perpendicular to the width direction of the calibration plate and perpendicular to the camera optical axis as a second coordinate axis to be calibrated, and marking the coordinate values ​​on the first coordinate axis to be calibrated as second coordinates; setting a coordinate axis parallel to the width direction of the calibration plate as a reference coordinate axis, and marking the coordinate values ​​on the reference coordinate axis as third coordinates, wherein the coordinates of the laser point in the width direction are the third coordinates; and then the plane formed by the intersection of the first coordinate axis to be calibrated and the reference coordinate axis, or the plane formed by the intersection of the second coordinate axis to be calibrated and the reference coordinate axis, can be set as a reference plane.

[0064] If the point cloud of an object in the calibration board has multiple planes or other features, it is necessary to set a certain plane to perform fitting analysis on the point cloud. In this embodiment, the plane where the coordinate points participating in the curve fitting are located or the plane on which the coordinate points participating in the curve fitting are uniformly projected can be set to be parallel to the reference plane.

[0065] In some embodiments, it is necessary to calibrate the distance between the camera and the object containing the laser image it acquires. In this case, the plane formed by the intersection of the first coordinate axis to be calibrated and the reference coordinate axis is set as the reference plane, and the point cloud of the calibration plane is projected onto the reference plane for curve fitting.

[0066] It should be noted that the first or second coordinate is designated as the coordinate to be calibrated, and correspondingly, the first or second coordinate axis to be calibrated is denoted as the calibration axis. The direction of the first coordinate axis to be calibrated represents the distance direction, and the direction of the second coordinate axis to be calibrated represents the height direction. Schematic, the first coordinate axis to be calibrated can be configured as the z-axis in a spatial coordinate system to measure distance information, and the reference coordinate axis can be configured as the x-axis in a spatial coordinate system, where the origin of the spatial coordinate system is the camera. If it is necessary to calibrate the height of the object in the laser image acquired by the camera, that is, to calibrate the height of the center of the laser stripes distributed column by column in the laser image, the plane formed by the intersection of the second coordinate axis to be calibrated and the reference coordinate axis is set as the reference plane. The point cloud of the calibration plane is projected onto this reference plane for curve fitting. The second coordinate axis to be calibrated can be configured as the y-axis in a spatial coordinate system to measure height information, and the reference axis can be configured as the x-axis in a spatial coordinate system.

[0067] It should be noted that the camera's optical axis is perpendicular to the calibration plate. Based on this perpendicularity, the orientation of the calibration plate is determined, and the width direction of the calibration plate is considered the extension direction of one of its edges, generally the side with the larger side length, typically extending from the left to the right of the calibration plate. The width direction of the calibration plate represents the arrangement direction of the laser stripes in the laser image; it can also be understood as the arrangement direction of the individual laser stripes pre-calibrated by the line laser module within the laser image.

[0068] The laser image is an image of laser stripes formed by the laser line reflected back from the calibration plate by the laser module. The laser points distributed within the calibration plane point cloud are the center points of the laser stripes in the laser image. Subsequently, the edge contour of the object to be calibrated (including the target object) can be fitted, reflecting the curvature trend of the currently distributed laser point cloud within the laser image. This can demonstrate the degree of distortion of the laser stripes in the calibration plate edge region and the degree of deformation of the laser line projected onto the calibration plate edge position. However, the actual edge position may be hidden. Nevertheless, the arrangement direction of the laser stripes in the laser image can be considered unchanged due to distortion. The center point of the light stripe can be calculated using the gray-scale centroid method. The gray-scale centroid method obtains the coordinates of the gray-scale centroid by searching a single column of pixels in the image and weighted averaging the gray values ​​of the pixels. In a portion (effective search pixel range) or the entire region of a single column of the laser image, the centroid position point calculated by the gray-scale centroid method is the center point of the laser stripe.

[0069] Therefore, in this embodiment, the establishment of the standard three-dimensional data coordinate system or the reference plane in the standard three-dimensional data coordinate system for the camera sampling laser points depends on the optical axis of the camera; based on the standard three-dimensional data coordinate system (which is also equivalent to the camera spatial coordinate system) composed of the first coordinate axis to be calibrated, the second coordinate axis to be calibrated, and the reference coordinate axis, the distortion of the pixel coordinate system in the laser image acquired from the calibration plate is evaluated (for example, represented by the deformation of the laser stripes projected onto the calibration plate), including the degree of distortion in the direction of the first coordinate axis to be calibrated and the degree of distortion in the direction of the second coordinate axis to be calibrated.

[0070] Preferably, the first coordinate axis to be calibrated is parallel to the horizontal plane, so that the first coordinate represents the horizontal distance of the center point of the corresponding laser stripe in the laser image relative to the camera, specifically the straight-line distance between the center point of one column of laser stripes and the optical center of the camera or the projected distance on the horizontal plane; when the reference plane is formed by the intersection of the first coordinate axis to be calibrated and the reference coordinate axis, the first coordinate can be reflected as the coordinate value in the reference plane and the spatial coordinate system in which it is located.

[0071] Preferably, each column of the laser image is parallel to the width direction of the laser stripes, and the second coordinate axis to be calibrated is perpendicular to the horizontal plane, so that the second coordinate represents the height of the center point of the corresponding laser stripe in the laser image relative to the horizontal plane, specifically the distance between the center point of one column of laser stripes and the optical center of the camera (the origin of the pixel coordinate system) in the direction of the second coordinate axis to be calibrated; when the reference plane is formed by the intersection of the second coordinate axis to be calibrated and the reference coordinate axis, the second coordinate reflects the coordinate value in the reference plane and its spatial coordinate system.

[0072] It is worth noting that the line laser module senses at least two laser points with unequal first coordinates within the same line segment in the laser image along the reference coordinate axis. Because of distortion at the horizontal edge of the camera lens, the same line segment in the laser image can be a curved segment in the horizontal plane. Therefore, the first coordinates (horizontal distance from the camera) of the two laser points sensed within the same curved segment (especially the two points near the edge of the calibration plate) are unequal. Furthermore, the line laser module senses at least two laser points with unequal second coordinates within the same line segment in the laser image along the reference coordinate axis. Because of distortion at the vertical edge of the camera lens, the same line segment in the laser image can be a curved segment. Therefore, the second coordinates (height difference relative to the camera) of the two laser points sensed within the same curved segment (especially the two points near the edge of the calibration plate) are unequal. In summary, this reflects the error caused by edge distortion of the laser image.

[0073] As one embodiment, in step B, the method of obtaining the equation of the curve to be calibrated by curve fitting using the calibration plane point cloud collected at the reference position includes: fitting the equation of the curve to be calibrated using the measurement value to be calibrated and the third coordinate of the calibration plane point cloud according to the least squares method, which serves as the equation of the curve to be calibrated at the reference position; alternatively, the coordinate points projected by the calibration plane point cloud onto the reference plane can be fitted to form the equation of the curve to be calibrated at the reference position according to the least squares method, wherein the highest power in the equation of the curve to be calibrated is greater than or equal to 2, so that the curve formed by the equation of the curve to be calibrated in the reference plane has at least one peak or at least one trough, which facilitates the acquisition of the maximum and minimum points in the equation of the curve to be calibrated to determine more accurate compensation information.

[0074] For each reference position, the collected calibration plane point cloud contains a measurement value to be calibrated and a third coordinate. Substituting multiple third coordinates into the equation of the curve to be calibrated, the corresponding measurement value to be calibrated is obtained. Then, the third coordinates and the measurement value to be calibrated form a discrete point sequence, which is then approximated by a polynomial function.

[0075] Schematic, the highest degree in the fitted equation of the calibration plane point cloud, whether acquired at the first or second reference position, is 4.

[0076] Taking horizontal distance as an example, the quartic polynomial equation f(x) = a1 + b1*x + c1*x is fitted using the first coordinate z1 and the third coordinate x1 from the calibration plane point cloud acquired by the camera from the calibration plate at the first reference position. 2 +d1*x 3 +e1*x 4 The equation for the curve to be calibrated at the first reference position is given, where a1 is a constant term, and the coefficients of the fourth-order polynomial equation are b1, c1, d1, and e1, respectively. This allows for fitting the distance variation curve of the calibration plane point cloud relative to the camera along the reference coordinate axis at the first reference position, describing the curvature trend of the calibration plane point cloud from the distance dimension, so that a corresponding symmetrical flip can be performed subsequently.

[0077] Additionally, the quartic polynomial equation f(x) = a² + b²x + c²x is fitted using the first coordinate z² and the third coordinate x² from the calibration plane point cloud acquired by the camera from the calibration plate at the second reference position. 2 +d2*x 3 +e2*x 4The equation for the curve to be calibrated at the second reference position is given, where a2 is a constant term, and the coefficients of the fourth-order polynomial equation are b2, c2, d2, and e2, respectively. This allows for fitting the distance variation curve of the calibration plane point cloud relative to the camera along the reference coordinate axis at the second reference position, describing the curvature trend of the calibration plane point cloud from the distance dimension, so that a corresponding symmetrical flip can be performed subsequently.

[0078] As one embodiment, the method for setting the target symmetry axis based on the extreme value of the equation of the curve to be calibrated includes: when the third coordinate is equal to 0, extracting the coordinate points on the reference plane from the laser image where the center point of the corresponding column of distributed laser stripes is located, and then recording the extracted coordinate points as alignment points; in this embodiment, the calibration measurement value of the center point where the third coordinate is equal to 0 can represent the distance information or height information of the camera lens center / camera optical center. Generally, the calibration measurement value of the center point where the third coordinate is equal to 0 is equal to the calibration measurement value of the alignment point.

[0079] The compensation starting value is set based on the positional relationship between the alignment point and the extreme point of the curve equation to be calibrated. Specifically, based on the positional relationship between the alignment point and the maximum point of the curve equation to be calibrated and / or the positional relationship between the alignment point and the minimum point of the curve equation to be calibrated, the measurement value to be calibrated at the extreme point of the curve equation to be calibrated can be set as the compensation starting value, or the measurement value to be calibrated at the alignment point can be set as the compensation starting value. This constitutes the starting position information for performing a symmetrical flip on the curve equation to be calibrated. Due to the positioning error of the center point of some laser stripes caused by camera edge distortion or light intensity changes, the compensation starting value is not necessarily equal to 0. Therefore, the starting position for performing a symmetrical flip on the curve equation to be calibrated is not necessarily fixed at the camera lens center / camera optical center, but it will adapt to the reverse compensation requirements of the calibration plane point cloud.

[0080] Based on this, in this embodiment, the straight line parallel to the reference coordinate axis and whose measurement value to be calibrated is the compensation starting value is set as the target axis of symmetry. Specifically, the target axis of symmetry is a symmetry axis set for the curve equation to be calibrated, parallel to the reference coordinate axis, and its coordinate offset relative to the origin of the spatial coordinate system (representing the camera lens center) in the direction of the coordinate axis to be calibrated (the first or second coordinate axis to be calibrated) is equal to the compensation starting value. Thus, the compensation starting value is used to determine the reference position for axial symmetry flipping of the curve equation to be calibrated on the coordinate axis to be calibrated, and then the target axis of symmetry is used to determine a more accurate symmetry compensation region, wherein the symmetry compensation region is the region reached by the curve equation to be calibrated through symmetrical flipping via the target axis of symmetry.

[0081] Based on the above embodiments, the method of using the target symmetry axis to control the axisymmetric flipping of the curve equation to be calibrated to obtain the compensation curve equation includes: controlling all coefficients in the curve equation to be calibrated to be configured as their opposites, controlling the constant terms and exponents of variables in the curve equation to be calibrated to remain unchanged, and controlling the compensation starting value to be subtracted from the end of the curve equation to be calibrated containing the constant term to obtain the compensation curve equation. At each reference position, the compensation curve equation is symmetrical to the curve equation to be calibrated about the target symmetry axis. The bending trend of the calibration plane point cloud is described from the direction symmetrical to the curve equation to be calibrated about the target symmetry axis. Then, step C uses a local region symmetrical to the waveform fitted by the curve equation to be calibrated and a scaling factor to compensate for the calibration measurement values ​​of each laser point in the point cloud to be calibrated.

[0082] Schematic, the highest degree in the equation of the curve to be calibrated is 4. Taking distance calibration as an example, the fourth-degree polynomial equation f1(x) = a1 + b1*x + c1*x is fitted at the first reference position using the first coordinate z1 and the third coordinate x1 of each laser point in the calibration plane point cloud. 2 +d1*x 3 +e1*x 4 And determine the target symmetry axis z = -z1', that is, the initial compensation value is z1'; then use the target symmetry axis to pair f1(x) = a1 + b1*x + c1*x 2 +d1*x 3 +e1*x 4 An axisymmetric flip is performed, thereby flipping the equation of the curve to be calibrated to the equation of the compensation curve at the first reference position. The equation of the compensation curve at the first reference position is: f1(x)=a1-z1'-b1*x-c1*x 2 -d1*x 3 -e1*x 4 .

[0083] Additionally, at the second reference position, a fourth-order polynomial equation f2(x) = a2 + b2*x + c2*x is fitted using the first coordinate z2 and the third coordinate x2 of each laser point in the calibration plane point cloud. 2 +d2*x 3 +e1*x 4 And determine the target symmetry axis z = -z2', that is, the initial compensation value is z2'; then use the target symmetry axis to pair f2(x) = a2 + b2*x + c2*x 2 +d2*x 3 +e1*x 4An axisymmetric flip is performed, thus transforming the equation of the curve to be calibrated into the equation of the compensation curve at the second reference position. The equation of the compensation curve at the second reference position is: f2(x)=a2-z2'-b2*x-c2*x 2 -d2*x 3 -e2*x 4 .

[0084] Since the variable exponents of the compensation curve equation and the curve equation to be calibrated remain unchanged and the coefficients of each term are transformed into opposite numbers, a symmetrical compensation difference is formed between the two. The compensation curve equation can be used to compensate for the scaled measurement values ​​to be calibrated in the point cloud to be calibrated.

[0085] As one embodiment, the method for setting the compensation starting value based on the positional relationship between the alignment point and the extreme points of the curve equation to be calibrated includes: calculating and obtaining the maximum and minimum points of the curve equation to be calibrated, wherein the coordinates to be calibrated for the maximum point of the curve equation to be calibrated are the maximum values, and the measured values ​​to be calibrated for the minimum points of the curve equation to be calibrated are the minimum values. The local extreme points corresponding to the curve peaks (local maximum values) formed by the curve equation to be calibrated in the coordinate plane are the maximum points, and the local extreme points corresponding to the curve troughs (local minimum values) formed by the curve equation to be calibrated in the coordinate plane are the minimum points. The curve equation to be calibrated is treated as a curve function, the third coordinate is the independent variable of the curve function, and the measured values ​​to be calibrated are the dependent variable or function values ​​of the curve function.

[0086] The distance between the nearest maximum point and the alignment point in the equation of the curve to be calibrated is recorded as the first alignment comparison value; and the distance between the nearest minimum point and the alignment point in the equation of the curve to be calibrated is recorded as the second alignment comparison value. In some curve fitting scenarios, the center points of each column of laser stripes in the laser image where the point cloud of the same calibration plane is located are fitted to the equation of the curve to be calibrated using the least squares method. The center point of the laser stripe with the third coordinate equal to 0 is directly regarded as the alignment point. Then, the distance between the alignment point and each extreme point of the equation of the curve to be calibrated is compared, that is, the first alignment comparison value related to the maximum point and the second alignment comparison value related to the minimum point are calculated respectively.

[0087] If the first alignment comparison value is greater than the second alignment comparison value, then the alignment point is determined to be relatively close to the minimum point of the curve to be calibrated. The calibration measurement value of the alignment point is relatively close to the minimum value of the equation of the curve to be calibrated. In this case, the calibration measurement value of the minimum point in the equation of the curve to be calibrated closest to the alignment point is set as the compensation starting value. At this time, the minimum point in the equation of the curve to be calibrated closest to the alignment point is configured as the compensation zero point, serving as the center point of the target symmetry axis. If the second alignment comparison value is greater than the first alignment comparison value, then the alignment point is determined to be relatively close to the maximum point of the equation of the curve to be calibrated. The calibration measurement value of the alignment point is relatively close to the maximum value of the equation of the curve to be calibrated. In this case, the calibration measurement value of the maximum point in the equation of the curve to be calibrated closest to the alignment point is set as the compensation starting value. At this time, the maximum point in the equation of the curve to be calibrated closest to the alignment point is configured as the compensation zero point, serving as the center of the target symmetry axis. The method achieves the setting of the measured value of the extreme point of the curve equation to be calibrated as the compensation starting value based on the positional relationship between the alignment point and the extreme point of the curve equation to be calibrated. This allows the compensation zero point to be obtained by aligning and comparing the maximum or minimum value, instead of directly using the aforementioned alignment point as the compensation zero point, thus overcoming the problem of inaccurate distance measurement of some point cloud center points.

[0088] In addition, if the first alignment comparison value is equal to the second alignment comparison value, the calibration coordinates of the alignment point are set as the compensation starting value. That is, after comparing the extreme points, the alignment point is allowed to be set as the compensation zero point, and the influence of edge position distortion can be ignored. Alternatively, if the first alignment comparison value is equal to the second alignment comparison value, the calibration measurement value of the maximum or minimum point in the curve equation to be calibrated that is closest to the alignment point is set as the compensation starting value. In this case, the aforementioned alignment point is not directly used as the compensation zero point. Instead, the maximum or minimum point is selected as the compensation zero point. The influence of edge position distortion cannot be ignored, and the influence of edge position distortion is often reflected in the change of the maximum or minimum point relative to the alignment point.

[0089] Therefore, it is possible to set the calibration coordinates of the alignment point to be calibrated as the compensation starting value when the influence of edge position distortion is ignored, and to select the calibration measurement value of the extreme point of the curve equation to be calibrated as the compensation starting value when the influence of edge position distortion cannot be ignored, thus overcoming the problem of inaccurate distance measurement of some laser stripe center points.

[0090] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent transformations or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A coupling compensation method based on laser measurement, wherein the coupling compensation method is applied to a line laser module, the line laser module comprising a line laser emitter and a camera; characterized in that, Coupling compensation methods include: Step A: Control the line laser emitter to emit a laser line towards the calibration board and control the calibration board to move relative to the line laser module; then execute step B; Step B: When the calibration plate moves at least two reference positions relative to the line laser module, whenever the calibration plate moves to a reference position, control the camera to acquire the calibration plane point cloud from the calibration plate; then use the calibration plane point cloud to perform curve fitting and axisymmetric flipping to obtain the compensation curve equation; then execute step C. Step C: Whenever the calibration board moves relative to the line laser module to a position to be calibrated, the camera is controlled to acquire the point cloud to be calibrated from the calibration board; the scaling factor is calculated using the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud; and based on the positional relationship between the point cloud to be calibrated and the reference position, the scaling factor and the compensation curve equation are combined to compensate the measurement value of the point cloud to be calibrated, so as to obtain the final measurement value of the point cloud to be calibrated.

2. The coupling compensation method according to claim 1, characterized in that, In step C, the method for compensating the measurement values ​​of the point cloud to be calibrated based on the positional relationship between the point cloud to be calibrated and the reference position, combined with the scaling factor and the compensation curve equation, to obtain the final measurement values ​​of the point cloud to be calibrated includes: Determine the positional relationship between a laser point in the point cloud to be calibrated and the target calibration interval; wherein, the target calibration interval is the region between the first and second calibrated measurements, such that the two ends of the target calibration interval are the first and second reference positions, respectively; the first calibrated measurement is less than the second calibrated measurement. If a laser point in the point cloud to be calibrated is determined to be outside the target calibration interval, the end with the smallest difference between the laser point and the laser point to be calibrated measurement value is selected from the target calibration interval as the target reference position. Then, the laser point to be calibrated measurement value is compensated by combining the scaling factor of the target reference position and the compensation curve equation at the target reference position to obtain the final measurement value of the laser point. If a laser point in the point cloud to be calibrated is determined to be within the target calibration range, then the scaling factor of the first reference position, the scaling factor of the second reference position, the compensation curve equation at the first reference position, and the compensation curve equation at the second reference position are combined to perform compensation calculation on the measurement value to be calibrated for the laser point, thereby obtaining the pre-calibration result of the laser point at the first reference position and the pre-calibration result of the laser point at the second reference position. Then, using the mean of the measurement values ​​to be calibrated in the calibration plane point cloud at the first reference position and the mean of the measurement values ​​to be calibrated in the calibration plane point cloud at the second reference position, a weighted average is calculated on the pre-calibration result of the laser point at the first reference position and the pre-calibration result of the laser point at the second reference position to obtain the final measurement value of the laser point. Wherein, the measurement value to be calibrated at the first reference position is the first measurement value to be calibrated, and the measurement value to be calibrated at the second reference position is the second measurement value to be calibrated; The reference positions that the calibration plate needs to move include a first reference position and a second reference position.

3. The coupling compensation method according to claim 2, characterized in that, During step C, if it is determined that the measurement value to be calibrated of a laser point in the point cloud to be calibrated is less than the first measurement value to be calibrated, then it is determined that the laser point in the point cloud to be calibrated is outside the target calibration range and the target reference position is determined to be the first reference position. During step C, if it is determined that the measurement value to be calibrated of a laser point in the point cloud to be calibrated is greater than the measurement value to be calibrated of the second laser point in the point cloud to be calibrated, then it is determined that the laser point in the point cloud to be calibrated is outside the target calibration range and the target reference position is determined to be the second reference position.

4. The coupling compensation method according to claim 3, characterized in that, The method for compensating the uncalibrated measurement value of a laser point by combining the scaling factor of the target reference position and the compensation curve equation at the target reference position to obtain the final measurement value of the laser point includes: Substitute the product of the coordinates of a laser point in the width direction of the point cloud to be calibrated and the lateral scaling factor of the target reference position into the compensation curve equation at the first reference position to obtain the pre-compensation result; then add the product of the pre-compensation result, the longitudinal scaling factor of the target reference position, and the amplitude scaling factor of the target reference position to the calibration measurement value of a laser point in the point cloud to be calibrated to obtain the final measurement value of a laser point in the point cloud to be calibrated. The scaling factors for the target reference position include the horizontal scaling factor, the vertical scaling factor, and the amplitude scaling factor.

5. The coupling compensation method according to claim 2, characterized in that, During step C, if the calibration measurement value of a laser point in the point cloud to be calibrated is greater than or equal to the first calibration measurement value, and the calibration measurement value of the same laser point is less than or equal to the second calibration measurement value, then it is determined that a laser point in the point cloud to be calibrated is located within the target calibration range.

6. The coupling compensation method according to claim 5, characterized in that, In the process of executing step C, by combining the scaling factor of the first reference position, the scaling factor of the second reference position, the compensation curve equation at the first reference position, and the compensation curve equation at the second reference position, the method for calculating the compensation of the laser point to be calibrated, and obtaining the laser point pre-calibration result at the first reference position and the laser point pre-calibration result at the second reference position, includes: Substitute the product of the coordinates of a laser point in the width direction of the point cloud to be calibrated and the first lateral scaling factor into the compensation curve equation at the first reference position to obtain the first pre-compensation result; then add the product of the first pre-compensation result, the first longitudinal scaling factor, and the first amplitude scaling factor to the calibration measurement value of a laser point in the point cloud to be calibrated to obtain the laser point pre-calibration result at the first reference position; wherein, the lateral scaling factor, the longitudinal scaling factor, and the amplitude scaling factor at the first reference position are denoted as the first lateral scaling factor, the first longitudinal scaling factor, and the first amplitude scaling factor, respectively. Substitute the product of the coordinates of a laser point in the width direction of the point cloud to be calibrated and the second lateral scaling factor into the compensation curve equation at the second reference position to obtain the second pre-compensation result; then add the product of the second pre-compensation result, the second longitudinal scaling factor, and the second amplitude scaling factor to the calibration measurement value of a laser point in the point cloud to be calibrated to obtain the laser point pre-calibration result at the second reference position; wherein, the lateral scaling factor, longitudinal scaling factor, and amplitude scaling factor at the second reference position are denoted as the second lateral scaling factor, the second longitudinal scaling factor, and the second amplitude scaling factor, respectively.

7. The coupling compensation method according to claim 6, characterized in that, The method for calculating a weighted average of the laser point pre-calibration results at the first reference position and the second reference position using the mean of the calibration plane point cloud to be calibrated and the mean of the calibration plane point cloud to be calibrated at the second reference position includes: The mean value of the calibration plane point cloud at the first reference position is recorded as the first reference amplitude mean value; and the absolute value of the difference between the calibration measurement value of a laser point in the calibration point cloud and the first reference amplitude mean value is recorded as the first relative compensation difference value. The mean value of the calibration plane point cloud at the second reference position is recorded as the second reference amplitude mean value; and the absolute value of the difference between the calibration measurement value of a laser point in the calibration point cloud and the second reference amplitude mean value is recorded as the second relative compensation difference value. The absolute value of the difference between the mean of the first reference amplitude and the mean of the second reference amplitude is denoted as the interval compensation difference. The ratio between the first relative compensation difference and the interval compensation difference is recorded as the second weight, and the ratio between the second relative compensation difference and the interval compensation difference is recorded as the first weight. The laser point pre-calibration result at the first reference position is multiplied by the second weight to obtain the first weighted value; The laser point pre-calibration result at the second reference position is multiplied by the first weight to obtain the second weighted value; Then, the first weighted value and the second weighted value are added together to obtain the final measurement value of a laser point in the point cloud to be calibrated.

8. The coupling compensation method according to claim 2, characterized in that, The method for calculating the scaling factor using the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud includes: When all laser points within the point cloud to be calibrated are distributed on the same plane, the following applies: Calculate the coordinate range of the point cloud to be calibrated in the width direction to obtain the current lateral range; where the width direction is the width direction of the laser stripes or the arrangement direction of the laser stripes pre-calibrated within the calibration plate; Calculate the range of the measurement values ​​to be calibrated in the point cloud to be calibrated, and obtain the current longitudinal range; Calculate the mean of the measurements to be calibrated in the point cloud to obtain the current mean amplitude; For a reference position, calculate the coordinate range of the calibration plane point cloud in the width direction to obtain the reference lateral range; then set the ratio between the current lateral range and the reference lateral range as the lateral scaling factor of the reference position. For a reference position, calculate the range of the measurement values ​​to be calibrated in the calibration plane point cloud to obtain the reference longitudinal range; then set the ratio between the current longitudinal range and the reference longitudinal range as the longitudinal scaling factor for the reference position. For a reference position, calculate the mean of the measurement values ​​to be calibrated in the calibration plane point cloud to obtain the reference amplitude mean; then set the ratio between the current amplitude mean and the reference amplitude mean as the amplitude scaling factor for the reference position. The scaling factors for the reference position include the horizontal scaling factor, the vertical scaling factor, and the amplitude scaling factor.

9. The coupling compensation method according to claim 8, characterized in that, The method for calculating the scaling factor using the proportional relationship between the point cloud to be calibrated and the calibration plane point cloud further includes: When the laser points within the point cloud to be calibrated are not distributed on the same plane, the following applies: At a reference position, the mean value of the measurement to be calibrated in the calibration plane point cloud is calculated. For each laser point in the point cloud to be calibrated, the ratio between the measurement to be calibrated of the laser point and the mean value of the measurement to be calibrated in the calibration plane point cloud is calculated. The calculated ratio is then set as the horizontal scaling factor, vertical scaling factor, or amplitude scaling factor of the reference position, where the horizontal scaling factor is equal to the vertical scaling factor and the vertical scaling factor is equal to the amplitude scaling factor.

10. The coupling compensation method according to claim 2, characterized in that, In step B, the method for obtaining the compensation curve equation by performing curve fitting and axisymmetric processing using the calibration plane point cloud at the reference position includes: Curve fitting is performed using the calibration plane point cloud collected at the reference location to obtain the equation of the curve to be calibrated; The target symmetry axis is set based on the extreme value of the curve equation to be calibrated, and then the target symmetry axis is used to control the curve equation to be calibrated to be axially symmetrically flipped to obtain the compensation curve equation at the reference position; wherein, the compensation curve equation at the reference position includes the compensation curve equation at the target reference position.

11. The coupling compensation method according to claim 10, characterized in that, The coupling compensation method further includes setting a coordinate axis parallel to the camera optical axis as a first coordinate axis to be calibrated, and marking the coordinate values ​​on the first coordinate axis to be calibrated as the first coordinate. The coordinate axis that is perpendicular to the width direction of the calibration plate and perpendicular to the optical axis of the camera is set as the second coordinate axis to be calibrated, and the coordinate values ​​on the second coordinate axis to be calibrated are marked as the second coordinate. The coordinate axis parallel to the width direction of the calibration plate is set as the reference coordinate axis, and the coordinate values ​​on the reference coordinate axis are marked as the third coordinate, where the coordinate of the laser point in the width direction is the third coordinate; The first coordinate or the second coordinate is designated as the measurement value to be calibrated; wherein the direction of the first coordinate axis to be calibrated is used to represent the distance direction, and the direction of the second coordinate axis to be calibrated is used to represent the height direction; The camera's optical axis is perpendicular to the calibration plate; the laser line is reflected back to the laser module through the calibration plate to form an image of laser stripes, and the laser points distributed in the calibration plane point cloud are represented by the center point of the laser stripes in the laser image.

12. The coupling compensation method according to claim 11, characterized in that, In step B, the method of obtaining the equation of the curve to be calibrated by curve fitting using the calibration plane point cloud collected at the reference position includes: The calibration curve equation is fitted using the least squares method with the measured values ​​to be calibrated and the third coordinate of the calibration plane point cloud, wherein the highest degree of the calibration curve equation is greater than or equal to 2.

13. The coupling compensation method according to claim 12, characterized in that, The method for setting the target symmetry axis based on the extreme values ​​of the equation of the curve to be calibrated includes: When the third coordinate is equal to 0, the center point of the corresponding column of laser stripes is extracted from the laser image where the calibration plane point cloud is located, and the extracted center point is recorded as the alignment point; Set the initial compensation value based on the positional relationship between the alignment point and the extreme point of the curve equation to be calibrated; The straight line whose measurement value to be calibrated is the compensation starting value and is parallel to the reference coordinate axis is set as the target axis of symmetry.

14. The coupling compensation method according to claim 13, characterized in that, The method of using the target axis of symmetry to control the equation of the curve to be calibrated to perform axisymmetric flipping to obtain the compensation curve equation at the reference position includes: The coefficients in the equation of the curve to be calibrated are configured to be their opposites, and the constant terms and exponents of the variables in the equation of the curve to be calibrated are kept constant. The compensation starting value is subtracted from the end of the equation of the curve to be calibrated containing the constant term to obtain the compensation curve equation at the reference position.

15. The coupling compensation method according to claim 13, characterized in that, The method for setting the initial compensation value based on the positional relationship between the alignment point and the extreme point of the curve equation to be calibrated includes: Calculate the maximum and minimum points of the equation of the curve to be calibrated. The distance between the nearest maximum point to the alignment point in the equation of the curve to be calibrated and the alignment point is recorded as the first alignment comparison value; The distance between the nearest minimum point in the equation of the curve to be calibrated and the alignment point is recorded as the second alignment comparison value; If the first alignment comparison value is greater than the second alignment comparison value, then the measurement value of the minimum point in the curve equation to be calibrated that is closest to the alignment point is set as the compensation starting value; If the second alignment comparison value is greater than the first alignment comparison value, then the measurement value of the maximum point in the curve equation to be calibrated that is closest to the alignment point is set as the compensation starting value; If the first alignment comparison value is equal to the second alignment comparison value, then the measurement value to be calibrated at the alignment point is set as the compensation starting value; or, if the first alignment comparison value is equal to the second alignment comparison value, then the measurement value to be calibrated at the nearest maximum or minimum point in the curve equation to be calibrated is set as the compensation starting value.

Images