Field of view angle range correction method, computer device and storage medium

Abstract

Embodiments of the present application provide a field of view angle range correction method, computer equipment and a storage medium. The method comprises: obtaining a target point cloud; determining a target inter-plate distance based on the target point cloud; obtaining a second point cloud set, determining a candidate inter-plate distance based on each frame of point cloud in the second point cloud set; and determining a target driving parameter based on the candidate inter-plate distance and the target inter-plate distance. The target point cloud obtained by the embodiments of the present application suppresses the point cloud distortion caused by the wide-angle lens, which is conducive to obtaining a reliable and accurate target inter-plate distance. The target inter-plate distance is used as a quantitative target for adjusting the driving parameter, and the driving parameter is constantly adjusted to obtain a target driving parameter that can make the candidate inter-plate distance consistent with the target inter-plate distance. The laser radar can drive the galvanometer to scan through the wide-angle lens at the target field of view angle range based on the target driving parameter. Therefore, the embodiments of the present application can help the laser radar to accurately and reliably output the laser line corresponding to the target field of view angle range.

Description

Technical Field

[0001] This application relates to the field of lidar technology, and in particular to a method for correcting the field of view, a computer device, and a storage medium. Background Technology

[0002] LiDAR systems are equipped with scanning devices, and the scanning angle of these devices affects the LiDAR's field of view (FOV). To expand the FOV, related technologies employ wide-angle lenses to magnify it. However, installation errors in the scanning devices and wide-angle lenses during practical applications can lead to significant differences between the actual and expected FOV, causing the LiDAR to fail to scan within the intended FOV range. Summary of the Invention

[0003] One objective of this application is to provide a method, computer device, and storage medium for correcting the field of view range, thereby improving the inaccuracy of the field of view range in related technologies.

[0004] In a first aspect, embodiments of this application provide a method for correcting the field of view, applied to a lidar. The lidar includes a galvanometer and an expanding lens, the galvanometer and the expanding lens being arranged opposite to each other. The correction method includes: acquiring a first point cloud set, wherein the first point cloud set includes multiple frames of point clouds obtained by the lidar driving the galvanometer to scan two calibration plates spaced apart using the same driving parameters; performing expanding distortion correction processing on the point clouds in the first point cloud set to obtain a target point cloud; determining the distance between the target plates based on the target point cloud; acquiring a second point cloud set, wherein the second point cloud set includes multiple frames of point clouds obtained by the lidar driving the galvanometer to scan the two calibration plates using different driving parameters; determining the distance between candidate plates based on each frame of point clouds in the second point cloud set; and determining target driving parameters based on the distance between candidate plates and the distance between the target plates, wherein the target driving parameters are used to drive the galvanometer to scan through the expanding lens within the target field of view.

[0005] This application embodiment performs angular distortion correction processing on the point cloud of the first point cloud set. The obtained target point cloud suppresses the point cloud distortion caused by the angular expansion lens, which is beneficial to obtain a more reliable and accurate target plate distance. The target plate distance is used as a quantization target for adjusting the driving parameters. The driving parameters are continuously adjusted to obtain target driving parameters that make the candidate plate distance consistent with the target plate distance. Based on the target driving parameters, the lidar can drive the galvanometer to scan through the angular expansion lens to the target field of view. Therefore, this application embodiment can help the lidar accurately and reliably output laser lines corresponding to the target field of view.

[0006] In some embodiments, performing expansion angle distortion correction processing based on the point cloud in the first point cloud set to obtain a target point cloud includes: generating an expansion angle calibration curve based on the first point cloud set, wherein the expansion angle calibration curve is used to correct the point cloud distortion generated by the expansion angle lens; and correcting the point cloud in the first point cloud set based on the expansion angle calibration curve to obtain the target point cloud.

[0007] In this embodiment, the point cloud in the first point cloud set is subjected to angular distortion correction processing. The resulting target point cloud compensates for the distortion caused by the addition of the angular correction lens, which facilitates subsequent steps to quickly, reliably and accurately obtain the target field of view based on the target point cloud.

[0008] In some embodiments, determining the target inter-plate distance based on the target point cloud includes: determining an adjustment ratio based on the target point cloud, the adjustment ratio being used to reflect the difference between the current field of view range of the lidar and a specified target field of view range; determining a reference inter-plate distance based on the first point cloud set, wherein the reference inter-plate distance is an inter-plate distance after angular distortion correction processing; and determining the target inter-plate distance based on the adjustment ratio and the reference inter-plate distance.

[0009] The embodiments of this application determine the adjustment ratio based on the target field of view range and the current field of view range, measure the difference between the current field of view range and the target field of view range, facilitate the subsequent derivation of the distance between target plates, provide a quantitative target for how to find accurate driving parameters, and thus accurately and reliably correct the current field of view range of the lidar to the target field of view range.

[0010] In some embodiments, determining the adjustment ratio based on the target point cloud includes: determining the current field of view of the lidar based on the target point cloud; and determining the adjustment ratio based on a preset target field of view and the current field of view.

[0011] In some embodiments, the target point cloud includes multiple rows of sub-point clouds, and determining the current field of view of the lidar based on the target point cloud includes: selecting sub-point clouds that meet preset position conditions from the target point cloud as target sub-point clouds; and determining the current field of view of the lidar based on the target sub-point clouds.

[0012] In some embodiments, the step of selecting a sub-point cloud that meets a preset position condition from the target point cloud as a target sub-point cloud includes: determining the middle row index of the target point cloud; determining that the sub-point cloud corresponding to the middle row index meets the preset position condition; and setting the sub-point cloud corresponding to the middle row index as the target sub-point cloud.

[0013] The actual point cloud acquired exhibits a shape that is "wider at the beginning and end, narrowing in the middle." Therefore, the horizontal field of view of the target sub-point cloud corresponding to the middle row index is the smallest. This embodiment selects the target sub-point cloud corresponding to the middle row index to determine the current field of view, which can effectively reflect the bottleneck of the lidar's field of view. After subsequently correcting the current field of view of the target sub-point cloud to the target field of view, the field of view of the sub-point clouds corresponding to other row indices will naturally be at least equal to or greater than the target field of view. This avoids the problem of not being able to detect objects due to an excessively small field of view, thus improving the accuracy and reliability of the lidar's field of view correction.

[0014] In some embodiments, the target sub-point cloud includes multiple point cloud points, and determining the current field of view of the lidar based on the target sub-point cloud includes: acquiring the coordinate information of each point cloud point; determining the yaw angle corresponding to the point cloud point based on the coordinate information of the point cloud point; searching for the minimum yaw angle and the maximum yaw angle among the yaw angles of the multiple point cloud points; and determining the current field of view of the lidar based on the minimum yaw angle and the maximum yaw angle.

[0015] In some embodiments, determining the distance between reference plates based on the first point cloud set includes: selecting a point cloud corresponding to the target attitude from the first point cloud set as a reference point cloud, wherein the laser line emitted by the lidar from the direction corresponding to the middle field of view under the target attitude is located in the middle of the two calibration plates; obtaining an expansion angle calibration curve; correcting the reference point cloud based on the expansion angle calibration curve to obtain a corrected reference point cloud; and determining the distance between reference plates based on the corrected reference point cloud.

[0016] The reference point cloud obtained by the lidar under the target attitude can more accurately and reliably represent the distance between the first calibration plate and the second calibration plate compared with other point clouds. In the embodiments of this application, after correcting the reference point cloud using the expansion angle calibration curve, a more accurate distance between the reference plates can be obtained based on the corrected reference point cloud.

[0017] In some embodiments, determining the target driving parameter based on the candidate board distance and the target board distance includes: calculating the absolute value of the difference between the candidate board distance and the target board distance; and determining the driving parameter corresponding to the candidate board distance as the target driving parameter in response to the absolute value of the difference being less than or equal to a preset threshold.

[0018] This application embodiment can find the target driving parameter with the smallest error by comparing the absolute value of the difference with a preset threshold. When the lidar works according to the target driving parameter, the output field of view range is consistent with the target field of view range or the error is very small, thereby achieving accurate correction of the field of view range.

[0019] In a second aspect, embodiments of this application provide a computer device including a memory and a processor. The memory is connected to the processor, and the processor is configured to execute one or more computer programs stored in the memory. When the processor executes the one or more computer programs, it causes the computer device to implement the above-described method for correcting the field of view range.

[0020] In a third aspect, embodiments of this application provide a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the aforementioned field-of-view range correction method. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A schematic diagram illustrating a scenario where an angle is expanded using an expanding lens, provided for related technologies;

[0023] Figure 2 A schematic diagram of the feature data points provided for related technologies in the three-dimensional coordinate system of lidar;

[0024] Figure 3 A schematic diagram of the expansion curve provided for related technologies;

[0025] Figure 4 A schematic flowchart illustrating a method for correcting the field of view range provided in an embodiment of this application;

[0026] Figure 5 A schematic diagram of a scenario where a turntable drives a lidar to scan a first calibration plate and a second calibration plate in a target posture, as provided in an embodiment of this application;

[0027] Figure 6 A schematic diagram of a scenario where a turntable drives a lidar to scan the first calibration plate and the second calibration plate clockwise, as provided in an embodiment of this application.

[0028] Figure 7 A schematic diagram of a scenario where a turntable drives a lidar to scan the first calibration plate and the second calibration plate counterclockwise, as provided in an embodiment of this application;

[0029] Figure 8 A schematic diagram of the structure of a field-of-view range correction device provided in an embodiment of this application;

[0030] Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0032] It should be noted that, unless there is a conflict, the various features in the embodiments of this application can be combined with each other, all of which are within the protection scope of this application. Furthermore, although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than the module division in the device or the order in the flowchart. Moreover, the terms "first," "second," and "third" used in this application do not limit the data or execution order, but only distinguish identical or similar items with essentially the same function and effect.

[0033] LiDAR scanning devices are divided into one-dimensional scanning devices and two-dimensional scanning devices. For example, a one-dimensional scanning device may include a rotating mirror, and a two-dimensional scanning device may include a MEMS galvanometer. Both one-dimensional and two-dimensional scanning devices are linear scanning devices, and the laser line is magnified according to the optical magnification after passing through the expanding lens.

[0034] Please see Figure 1 The MEMS galvanometer 11 and the expanding lens 12 are positioned opposite each other, with the center of the MEMS galvanometer 11 and the center of the expanding lens 12 on the same horizontal line. After the laser line 13 enters the MEMS galvanometer 11, it is reflected by the MEMS galvanometer 11 to the expanding lens 12. The expanding lens 12 magnifies the exit angle of the laser line 13 according to a preset optical magnification. Related technologies, combined with a preset expanding angle model, correct the point cloud after expanding the angle. The expanding angle model is represented by an expanding angle curve. The expanding angle curve in related technologies is a design curve; due to manufacturing tolerances, the actual expanding angle curve of the expanding lens often differs from the theoretical expanding angle curve. Therefore, when the lidar uses the theoretical expanding angle curve to process the point cloud, the resulting point cloud is prone to distortion.

[0035] The embodiments of this application relate to the principle of angle expansion distortion correction. To facilitate understanding of the embodiments of this application, the principle of angle expansion distortion correction is first explained as follows:

[0036] Please see Figure 2The formula for calculating the coordinates (x, y, z) of feature data points in the point cloud before expansion is as follows:

[0037]

[0038] (Formula 1)

[0039]

[0040] in,( , ) is the outgoing vector, and dist is the target distance corresponding to the feature data point.

[0041] The lidar is configured with a three-dimensional coordinate system XYZ. The X-axis corresponds to the front-back direction of the lidar, the Y-axis corresponds to the left-right direction of the lidar, and the Z-axis corresponds to the vertical direction of the lidar. r is the distance dist from the feature data point to the origin of the three-dimensional coordinate system XYZ.

[0042] The outgoing vector Convert to unit spherical coordinates:

[0043]

[0044] (Formula 2)

[0045]

[0046] in, Let be the tangential angle of the outgoing vector (i.e., the angle without expansion), that is: The angle between the direction of the laser beam reflected by the galvanometer and the optical axis of the expanding lens. The angle between the normals of the outgoing vectors is... This is the azimuth angle in a three-dimensional coordinate system. It's understandable that, since the expanding lens only magnifies the tangential angle of the laser line in the tangential direction and not the radial normal angle, the normal angle remains unchanged.

[0047] The theoretical curve for an extended-angle lens is:

[0048] (Formula 3)

[0049] in, The coefficients of the fourth-order term in a fourth-order polynomial. The coefficients of the cubic terms in a fourth-order polynomial. The coefficients of the quadratic term in the fourth-order polynomial. The coefficients of the linear terms in a fourth-order polynomial. The coefficient of the zeroth term of the fourth-order polynomial. This is the first expansion angle after expansion.

[0050] Please see Figure 3 The lidar obtains the angle before the expansion angle. Afterwards, based on Figure 3 The expansion curve shown can be used to obtain the first expansion angle after expansion. Based on Formula 2, the expanded emission vector is obtained. , , ), as shown below:

[0051]

[0052] (Formula 4)

[0053]

[0054] As can be seen from Formula 4 and Formula 2, the included normal angle remains unchanged.

[0055] For the same feature data point, the distance before and after the expansion angle are equal. Using formulas 1 and 4, the coordinates of the corrected feature data point can be derived. , , The corrected feature data points together form the corrected point cloud. This concludes the introduction to the principle of angular distortion correction.

[0056] The expansion angle model provided by related technologies has been widely applied to multiple expansion angle lenses. However, it is difficult to ensure consistency in manufacturing processes and installation precision among different expansion angle lenses. Manufacturing errors and individual differences exist between different lenses. When the same expansion angle model is applied to different lenses, even after expansion angle correction using the expansion angle model, the point clouds corresponding to different lenses still have errors and require recalibration. It is understandable that the LiDAR is equipped with an expansion angle lens before leaving the factory; therefore, the recalibration process of the expansion angle curve is actually based on the expanded angle obtained from point 2. It is worth noting that, in this embodiment of the application, This refers to the angle before the expansion angle. This refers to the angle after expansion. This refers to the angle obtained by calibrating the angle after expansion.

[0057] The scanning angle of the galvanometer is linearly related to the driving signal, while the magnification of the expanding lens is non-linearly related to the incident angle. When the lidar linearly adjusts the scanning range of the galvanometer, the output field of view of the lidar changes non-linearly under the combined effect of the galvanometer and the expanding lens. Installation errors in the galvanometer and expanding lens in practical applications can lead to a significant difference between the actual and expected field of view, causing the lidar to fail to scan within the expected field of view.

[0058] Related technologies can measure the angles of two edge field-of-view angles of a lidar, and determine the current field-of-view range based on these angles. When the current field-of-view range does not reach the specified target field-of-view range, these technologies need to continuously adjust the drive of the galvanometer to correct the current field-of-view range to the specified target field-of-view range. However, when an extended-angle lens is added to the lidar, the output field-of-view range of the lidar changes non-linearly. It is difficult for these technologies to find a drive that makes the current field-of-view range equal to the target field-of-view range by continuously adjusting the galvanometer drive. This approach suffers from low correction efficiency, and the final corrected field-of-view range still differs significantly from the target field-of-view range.

[0059] The embodiments of this application can correct the point cloud distortion generated by the wide-angle lens, and then correct the field of view of the lidar based on this. This helps to improve the accuracy of the field of view correction and can quickly find the target driving parameters that can correct the current field of view to the target field of view, thereby improving the correction efficiency.

[0060] The following embodiments of this application provide a method for correcting the field of view range. Please refer to... Figure 4 The embodiments of this application implement a method for correcting the field of view range through steps S41 to S46, as detailed below:

[0061] Step S41: Obtain the first point cloud set.

[0062] The first point cloud set comprises multiple frames of point clouds obtained by scanning two calibration plates spaced apart using a lidar-driven galvanometer with the same driving parameters. The galvanometer can be a one-dimensional or two-dimensional galvanometer, and the driving parameters are those that drive the galvanometer's rotation. These parameters can be the driving current or driving voltage, and can be linear, sinusoidal, sawtooth, or stepped, or a combination of linear and nonlinear driving parameters.

[0063] The point cloud is obtained by scanning two calibration boards with a lidar in a corresponding posture. The lidar outputs laser lines within its field of view in that posture to scan the two calibration boards. The field of view is the range of angles that the lidar can scan. It can be understood that the field of view includes the horizontal field of view and the vertical field of view. The horizontal field of view is the range that the lidar can scan horizontally, and the vertical field of view is the range that the lidar can scan vertically. For example, the user requires the lidar to have a horizontal field of view of [0°, 120°] and a vertical field of view of [0°, 60°].

[0064] Two calibration plates are spaced apart at the calibration location. Please refer to [link / reference]. Figure 5 In this embodiment of the application, a calibration site is set up, in which a first calibration plate 31, a second calibration plate 32, a turntable 33 and a lidar 34 are set up.

[0065] The first calibration plate 31 and the second calibration plate 32 are respectively arranged on opposite sides of the lidar 34. The first calibration plate 31 and the second calibration plate 32 are spaced apart and parallel.

[0066] The lidar 34 is placed on the turntable 33, which drives the lidar 34 to rotate, changing the attitude of the lidar 34 so that the lidar 34 emits laser lines at different positions to scan the first calibration plate 31 and the second calibration plate 32.

[0067] The lidar 34 is equipped with a three-dimensional coordinate system XYZ. The X-axis is perpendicular to the emission surface of the lidar 34, the Y-axis is perpendicular to the X-axis in the horizontal plane, and the Z-axis is perpendicular to the Y-axis in the vertical plane. The first calibration plate 31 and the second calibration plate 32 are both perpendicular to the aforementioned Y-axis. Therefore, their y-coordinates in the three-dimensional coordinate system XYZ and the y-coordinates of the second calibration plate 32 in the three-dimensional coordinate system XYZ are both constants. The distance from the first calibration plate 31 or the second calibration plate 32 to the origin of the three-dimensional coordinate system XYZ is less than 20m, and the distance between the first calibration plate 31 and the second calibration plate 32 is less than 15m but greater than 3.5m.

[0068] Each time the turntable 33 changes the attitude of the lidar 34, the lidar 34 scans the first calibration plate 31 and the second calibration plate 32 in that attitude, thereby obtaining a point cloud frame. Similarly, when the turntable 33 changes the attitude of the lidar 34 multiple times, multiple point clouds can be obtained, and the multiple point clouds form a point cloud set.

[0069] For example, the turntable 33 starts to drive the lidar 34 to rotate according to a preset step angle, wherein the preset step angle is 1°.

[0070] Please continue reading. Figure 5First, in the initial state, the laser line 51 emitted from the direction corresponding to the middle field of view in the horizontal field of view range is located between the first calibration plate 31 and the second calibration plate 32. The lidar 34 scans the first calibration plate 31 and the second calibration plate 32 in the current posture to obtain the first point cloud F1.

[0071] Next, please refer to Figure 6 When the turntable 33 drives the lidar 34 to rotate clockwise by 1°, the lidar scans the first calibration plate 31 and the second calibration plate 32 in the current posture to obtain the second point cloud F2.

[0072] Next, please continue reading. Figure 6 When the turntable 33 drives the lidar 34 to rotate clockwise by 1°, the lidar scans the first calibration plate 31 and the second calibration plate 32 in the current posture to obtain the third point cloud F3.

[0073] This process continues until the laser line at the leftmost edge of the horizontal field of view just hits the second calibration plate 32. At this point, the lidar scans the first calibration plate 31 and the second calibration plate 32 in the current posture to obtain the m-th point cloud Fm.

[0074] Next, please refer to Figure 7 When the turntable 33 drives the lidar 34 to rotate counterclockwise by 1°, the lidar scans the first calibration plate 31 and the second calibration plate 32 in the current posture to obtain the (m+1)th point cloud Fm+1.

[0075] Next, please continue reading. Figure 7 When the turntable 33 drives the lidar 34 to rotate counterclockwise by 1°, the lidar scans the first calibration plate 31 and the second calibration plate 32 in the current posture to obtain the (m+2)th point cloud Fm+2.

[0076] This process continues until the laser line at the rightmost edge of the horizontal field of view just hits the first calibration plate 31. Then, the lidar scans the first calibration plate 31 and the second calibration plate 32 in the current posture to obtain the nth point cloud Fn.

[0077] Therefore, the first point cloud set is {F0,F1,F2,F3,...,Fm,Fm+1,...,Fn}. In this embodiment, the lidar is controlled to scan the first calibration plate 31 and the second calibration plate 32 using a full field-of-view method, ensuring that each posture corresponds to a point cloud. This allows for accurate calibration of the angle of the laser line in that posture before expansion, thereby obtaining a reliable, accurate, and comprehensive expansion calibration curve covering the entire field of view.

[0078] Step S42: Perform angular distortion correction processing on the point clouds in the first point cloud set to obtain the target point cloud.

[0079] The expansion distortion correction process corrects point cloud distortion caused by installation errors between the galvanometer and the expansion lens. In this embodiment, any frame of point cloud from the first point cloud set can be selected for expansion distortion correction, or a point cloud corresponding to the target attitude can be selected from the first point cloud set. For example, the target attitude is the attitude when the laser line emitted by the lidar from the direction corresponding to the middle field of view hits between the first and second calibration plates. In this embodiment, the point cloud acquired by the lidar under this target attitude is selected for expansion distortion correction.

[0080] In this embodiment of the application, the target point cloud is obtained by performing angular distortion correction processing based on the point cloud in the first point cloud set through steps S421 and S422, as shown below:

[0081] Step S421: Generate an expansion angle calibration curve based on the first point cloud set, wherein the expansion angle calibration curve is used to correct the point cloud distortion caused by the expansion angle lens.

[0082] In this embodiment, a first feature region corresponding to the first calibration board and / or a second feature region corresponding to the second calibration board are extracted from each frame of point cloud in the first point cloud set. Both the first and second feature regions contain multiple feature data points. A turntable rotates the LiDAR according to a preset step angle, and the LiDAR acquires one frame of point cloud data with each rotation. The LiDAR extracts the first feature region belonging to the first calibration board and / or the second feature region belonging to the second calibration board from the point cloud according to a preset wall recognition algorithm. The preset wall recognition algorithm includes deep learning algorithms, edge detection algorithms, or image analysis algorithms, etc.

[0083] The embodiments of this application determine the first expansion angle based on the feature data points and the first association relationship, whereby the first association relationship is the association relationship between the target distance of the feature data points and the three-dimensional coordinates.

[0084] The feature data point includes three-dimensional coordinates (x, y, z) and a target distance dist. The three-dimensional coordinates (x, y, z) are the coordinates of the feature data point in the three-dimensional coordinate system, and the target distance dist is the distance from the feature data point to the origin of the three-dimensional coordinate system. The first association relationship is used to describe the mathematical relationship between the three-dimensional coordinates (x, y, z) of the feature data point in the three-dimensional coordinate system and the target distance. In this embodiment, Formula 5 is selected to represent the first association relationship, as shown below:

[0085]

[0086] (Formula 5)

[0087]

[0088] Specifically, in this embodiment of the application, the first outgoing vector after expansion is determined based on the three-dimensional coordinates of the feature data points, the target distance, and the first correlation relationship. Based on the first outgoing vector and the second correlation relationship, the first expansion angle is determined. The second correlation relationship is used to represent the relationship between the first expansion angle and the first outgoing vector.

[0089] In this embodiment, the first outgoing vector after the expansion angle can be obtained using Formula 5. , , The second association relationship is represented by Formula 6, as follows:

[0090]

[0091] (Formula 6)

[0092]

[0093] in, This is the first expansion angle. Let be the included angle of the normals, where is the included angle of the normals. Let be the angle between the projection of the outgoing vector onto the normal plane and the Y-axis. As shown in Formula 6, the first expansion angle can be obtained in this embodiment based on the first outgoing vector and the second correlation relationship. Angle with normal .

[0094] The embodiments of this application determine the first angle before expansion based on the first expansion angle and the first expansion curve. The first expansion curve is the correlation curve between the first angle and the first expansion angle.

[0095] The first expansion curve is constrained by multiple expansion coefficients. From Formula 3, the expression for the first expansion curve is: .in, , , , and All are expansion angle coefficients. This is the first angle before the expansion angle.

[0096] This application embodiment determines the first angle before expansion based on a first expansion angle and multiple expansion coefficients. This application embodiment has obtained the first expansion angle using Formula 6. The first angle before expansion can be obtained using Formula 3. .

[0097] Next, the embodiments of this application determine the three-dimensional correction coordinates corresponding to the feature data points, and determine the second expansion angle based on the three-dimensional correction coordinates corresponding to the feature data points and the first correlation relationship.

[0098] The three-dimensional calibration coordinates are the coordinates obtained after performing an expansion distortion correction operation on the three-dimensional coordinates of the feature data points. The three-dimensional calibration coordinates reflect the true position of the feature data points. The expansion calibration curve is determined based on different first angles and corresponding second expansion angles.

[0099] Determining the three-dimensional calibration coordinates corresponding to feature data points includes the following steps: obtaining the ground truth data corresponding to the feature data points; and calibrating the three-dimensional coordinates of the feature data points based on the three-dimensional coordinates of the feature data points, the target distance, and the ground truth data to obtain the three-dimensional calibration coordinates of the feature data points.

[0100] The true value data is the data obtained by the high-precision lidar measuring the calibration plate. The true value data reflects the "true value" or "reference value" of the calibration plate under ideal conditions. In some embodiments, with the X-axis of the three-dimensional coordinate system XYZ perpendicular to the output surface of the lidar, the Y-axis perpendicular to the X-axis in the horizontal plane, and the Z-axis perpendicular to the Y-axis in the vertical plane, the true value data is the true value of the y-axis of the calibration plate on the Y-axis, that is: the true value of the y-axis of all feature data points is known.

[0101] Based on the property that the included normal angle remains unchanged before and after the expansion angle, the ratio between the y-coordinate and z-coordinate of the feature data point before correction and the ratio after correction are... coordinates and The ratio between them remains constant, that is, the following relationship holds:

[0102] (Formula 7)

[0103] Correcting the three-dimensional coordinates of feature data points to obtain their corrected three-dimensional coordinates involves the following steps: based on the three-dimensional coordinates and ground truth data of the feature data points, correcting the position information of the feature data points on the normal plane to obtain two-dimensional coordinates; determining the third-dimensional coordinates based on the two-dimensional coordinates and the target distance; and determining the three-dimensional corrected coordinates of the feature data points based on the two-dimensional coordinates and the third-dimensional coordinates.

[0104] In a three-dimensional coordinate system XYZ, where the X-axis is perpendicular to the laser radar's emission surface, the Y-axis is perpendicular to the X-axis in a horizontal plane, and the Z-axis is perpendicular to the Y-axis in a vertical plane, the normal plane is the plane formed by the Y-axis and the Z-axis. Equation 7 can be rewritten as Equation 8, as shown below:

[0105] (Formula 8)

[0106] The lidar has obtained the three-dimensional coordinates (x, y, z) of the feature data points, and the ground truth data. The value is known.

[0107] As can be seen from Formula 8, the embodiments of this application are based on the y-axis coordinate (y) and z-axis coordinate (z) of the feature data points on the Y-axis and the true value of the y-axis (…). Determine the z-axis correction coordinates. Based on the true value of the y-axis ( ) and z-axis correction coordinates ( ), to obtain two-dimensional coordinates ( , ).

[0108] The relationship between the target distance dist and the three-dimensional coordinates (x, y, z) is expressed by Equation 9, as shown below:

[0109] (Formula 9)

[0110] Formula 9 can be rewritten as Formula 10, as shown below:

[0111] (Formula 10)

[0112] As can be seen from Formula 10, in this embodiment of the application, the x-axis corrected coordinates of the feature data points are determined based on the true y-axis value, the z-axis corrected coordinates, and the target distance. x-axis correction coordinates This is the third-dimensional coordinate.

[0113] This application embodiment uses two-dimensional coordinates ( , ) and third-dimensional coordinates By combining the data, the three-dimensional corrected coordinates of the feature data points are obtained. , , ).

[0114] In some embodiments, the present application can directly determine the second expansion angle using the correlation between the three-dimensional correction coordinates of the feature data points and the first correlation relationship. In other embodiments, the present application can obtain the normal angle and then determine the second expansion angle based on the three-dimensional correction coordinates of the feature data points, the normal angle, and the first correlation relationship.

[0115] The determination of the second expansion angle based on the three-dimensional correction coordinates corresponding to the feature data points and the first correlation relationship includes the following steps: based on the three-dimensional correction coordinates of the feature data points, the target distance and the first correlation relationship, determine the second outgoing vector after expansion; based on the second outgoing vector and the second correlation relationship, determine the second expansion angle, where the second correlation relationship is the relationship between the second expansion angle and the second outgoing vector.

[0116] In this embodiment, the second emission vector after widening can be obtained from formula 11. , , ), as shown below:

[0117]

[0118] (Formula 11)

[0119]

[0120] In this application embodiment, Formula 12 is used to represent the second association relationship, as follows:

[0121]

[0122] (Formula 12)

[0123]

[0124] This is the second expansion angle. The angle between the normal and the normal. The second outgoing vector ( , , Given the given value, the second expansion angle can be obtained using Formula 12. .

[0125] Thus far, the embodiments of this application have obtained each directional angle. Corresponding feature data points Second expansion angle and the first angle The expression for the expansion angle description information of feature data point i is ( , , , The set of multiple feature data points under different orientation angles is {( , , , ),( , , , ),...,( , , , )}.

[0126] Based on different first angles and corresponding second expansion angles, the embodiments of this application reconstruct the expression of the expansion angle calibration curve, as shown in Formula 13:

[0127] (Formula 13)

[0128] This application embodiment uses a preset fitting algorithm to combine different first angles and corresponding second expansion angles to obtain the new expansion coefficient of Formula 13. , , , and For example, the preset fitting algorithm is the least squares method. This embodiment of the application uses the least squares method, which is used to fit {( , ),( , ),...,( , Substituting into formula 13, the new expansion coefficient is obtained by fitting. , , , and Based on the new expansion coefficient , , , and This allows us to obtain the mapping relationship between any first angle and its corresponding second expansion angle, thus obtaining the expansion angle calibration curve.

[0129] The expansion angle calibration curve is used to represent the mapping relationship between any first angle and second expansion angle. The expansion angle calibration curve can not only improve the expansion angle distortion caused by the introduction of expansion angle lens in lidar, but also further correct the expansion angle distortion caused by individual differences in expansion angle lens. This will enable lidar to obtain more accurate and reliable point clouds.

[0130] Step S422: Correct the point cloud in the first point cloud set based on the expansion angle calibration curve to obtain the target point cloud.

[0131] The target point cloud is a point cloud that has undergone angular distortion correction processing. The point cloud includes multiple point cloud points. In this embodiment, a first angle is determined for each point cloud point. The first angle is substituted into the expression of the angular distortion calibration curve to obtain the second angular distortion angle after angular distortion correction. Based on the second angular distortion angle after angular distortion correction, combined with formulas 11 and 12, the coordinates of the point cloud points after angular distortion correction processing are obtained. The coordinates of all point cloud points after angular distortion correction processing are combined to obtain the target point cloud.

[0132] In this embodiment, the point cloud in the first point cloud set is subjected to angular distortion correction processing. The resulting target point cloud compensates for the distortion caused by the addition of the angular correction lens, which facilitates subsequent steps to quickly, reliably and accurately obtain the target field of view based on the target point cloud.

[0133] Step S43: Determine the distance between target boards based on the target point cloud.

[0134] The embodiments of this application set the distance between target plates to provide a quantitative target for adjusting the driving parameters of the lidar, which enables the current field of view of the lidar to be corrected to the target field of view.

[0135] In this embodiment of the application, the distance between target boards is determined based on the target point cloud through steps S431 to S433, as detailed below:

[0136] Step S431: Determine the adjustment ratio based on the target point cloud.

[0137] The adjustment ratio is used to reflect the difference between the current field of view of the lidar and the specified target field of view. Determining the adjustment ratio based on the target point cloud includes the following steps: determining the current field of view of the lidar based on the target point cloud, and determining the adjustment ratio based on the preset target field of view and the current field of view.

[0138] The target field of view is customized by the designer according to product requirements. For example, the current field of view range... The target field of view range is [0°, 115°]. The range is [0°, 120°], and the adjustment ratio is... For another example, the current field of view range The target field of view range is [0°, 125°]. The range is [0°, 120°], and the adjustment ratio is... .

[0139] The embodiments of this application determine the adjustment ratio based on the target field of view range and the current field of view range, measure the difference between the current field of view range and the target field of view range, facilitate the subsequent derivation of the distance between target plates, provide a quantitative target for how to find accurate driving parameters, and thus accurately and reliably correct the current field of view range of the lidar to the target field of view range.

[0140] The target point cloud comprises multiple rows of sub-point clouds, each row of which is configured with a row index. The row index indicates the order of the sub-point cloud within the target point cloud, and the row indices of each row of sub-point clouds are arranged in a preset order. For example, the target point cloud comprises 101 rows of sub-point clouds, with the row index of the first row of sub-point clouds being 1, the row index of the second row of sub-point clouds being 2, and so on.

[0141] In some embodiments, determining the current field of view of the lidar based on the target point cloud includes the following steps: determining the field of view range corresponding to each row of sub-point clouds, performing average calculation on the field of view ranges corresponding to all rows of sub-point clouds to obtain the average field of view range, and using the average field of view range as the current field of view range of the lidar.

[0142] In other embodiments, embodiments of this application may select the field of view range corresponding to a sub-point cloud at a specified location as the current field of view range of the lidar. Specifically, determining the current field of view range of the lidar based on the target point cloud includes the following steps: selecting sub-point clouds that meet preset position conditions from the target point cloud as target sub-point clouds, and determining the current field of view range of the lidar based on the target sub-point clouds.

[0143] The preset position condition is the condition for selecting target sub-point clouds from the target point cloud. The components of the preset position condition include that the row index of the target sub-point cloud is the middle row index. Selecting the sub-point cloud that meets the preset position condition from the target point cloud as the target sub-point cloud includes the following steps: determining the middle row index of the target point cloud, determining that the sub-point cloud corresponding to the middle row index meets the preset position condition, and setting the sub-point cloud corresponding to the middle row index as the target sub-point cloud.

[0144] The middle row index is located in the middle position among all row indices in the target point cloud. When the total number of row indices N in the target point cloud is odd, the middle row index is... When the total number of row indices N of the target point cloud is odd, the middle row index is... For example, N=101, and the middle row index is 51. For example, N=100, and the middle row index is 50.

[0145] The actual point cloud acquired exhibits a shape that is "wider at the beginning and end, narrowing in the middle." Therefore, the horizontal field of view of the target sub-point cloud corresponding to the middle row index is the smallest. This embodiment selects the target sub-point cloud corresponding to the middle row index to determine the current field of view, which can effectively reflect the bottleneck of the lidar's field of view. After subsequently correcting the current field of view of the target sub-point cloud to the target field of view, the field of view of the sub-point clouds corresponding to other row indices will naturally be at least equal to or greater than the target field of view. This avoids the problem of not being able to detect objects due to an excessively small field of view, thus improving the accuracy and reliability of the lidar's field of view correction.

[0146] The target sub-point cloud includes multiple point cloud points. Determining the current field of view of the lidar based on the target sub-point cloud includes the following steps: acquiring the coordinate information of each point cloud point, determining the yaw angle corresponding to the point cloud point based on the coordinate information of the point cloud point, searching for the minimum yaw angle and the maximum yaw angle among the yaw angles of multiple point cloud points, and determining the current field of view of the lidar based on the minimum yaw angle and the maximum yaw angle.

[0147] The coordinates of each point cloud point in the lidar coordinate system are (x, y, z). In this embodiment, the point cloud points in the lidar coordinate system are transformed to the Cartesian coordinate system, as follows:

[0148] (Formula 14)

[0149] Yaw angle It is the pitch angle.

[0150] Based on Formula 14, this embodiment of the application can obtain the yaw angle and pitch angle of each point in the target sub-point cloud. This embodiment of the application searches for the minimum yaw angle within the target sub-point cloud. With maximum yaw angle The current field of view of the lidar is determined based on the minimum and maximum yaw angles. .

[0151] Step S432: Determine the distance between reference boards based on the first point cloud set.

[0152] The reference inter-board distance is the inter-board distance after angular distortion correction.

[0153] In some embodiments, determining the reference inter-board distance based on the first point cloud set includes the following steps: obtaining an expansion angle calibration curve, correcting each frame of the point cloud in the first point cloud set based on the expansion angle calibration curve to obtain a corrected point cloud, determining the first inter-board distance based on the corrected point cloud, performing an average value calculation operation on the first inter-board distance corresponding to each corrected point cloud to obtain an average inter-board distance, and setting the average inter-board distance as the reference inter-board distance.

[0154] In other embodiments, determining the distance between reference plates based on the first point cloud set includes the following steps: selecting a point cloud corresponding to the target attitude as a reference point cloud from the first point cloud set, wherein the laser line emitted by the lidar from the direction corresponding to the middle field of view under the target attitude is located in the middle of the two calibration plates, obtaining the expansion angle calibration curve, correcting the reference point cloud based on the expansion angle calibration curve, obtaining the corrected reference point cloud, and determining the distance between reference plates based on the corrected reference point cloud.

[0155] like Figure 5As shown, the lidar emits a laser line at the target's attitude. The laser line 51 emitted from the direction corresponding to the middle field of view within the horizontal field of view range is located between the first calibration plate 31 and the second calibration plate 32, resulting in a first point cloud F1. In this embodiment, the first point cloud F1 is set as a reference point cloud. Then, following the methods described in the above embodiments, the reference point cloud is corrected based on the expansion calibration curve to obtain a corrected reference point cloud. Next, this embodiment extracts a first feature region corresponding to the first calibration plate and a second feature region corresponding to the second calibration plate from the corrected reference point cloud, calculates the distance between the first and second feature regions, and uses this distance as the distance between the reference plates.

[0156] The reference point cloud obtained by the lidar under the target attitude can more accurately and reliably represent the distance between the first calibration plate and the second calibration plate compared with other point clouds. In the embodiments of this application, after correcting the reference point cloud using the expansion angle calibration curve, a more accurate distance between the reference plates can be obtained based on the corrected reference point cloud.

[0157] Step S433: Determine the target inter-board distance based on the adjustment ratio and the distance between the reference boards.

[0158] In this embodiment, the target inter-board distance is obtained by dividing the reference inter-board distance by an adjustment ratio, as shown below:

[0159] (Formula 15)

[0160] in, The target distance between boards. For reference, the distance between the boards To adjust the ratio.

[0161] Step S44: Obtain the second point cloud set.

[0162] The second point cloud set includes multiple frames of point clouds obtained by the lidar using different driving parameters to drive the galvanometer to scan two calibration plates. In this embodiment, the driving parameters are adjusted according to a preset gain step value, and the lidar is controlled to scan the two calibration plates based on the adjusted driving parameters to obtain point clouds. All point clouds are then combined to obtain the second point cloud set.

[0163] The gain step value is customized by the designer based on engineering experience; for example, the gain step value is 0.01. This application embodiment obtains the driving parameter range and the number of acquisitions. The driving parameter range is jointly defined by the minimum driving parameter and the maximum driving parameter. Based on the minimum driving parameter, the gain step value, and the number of acquisitions within the driving parameter range, the current driving parameter is determined. When the current driving parameter is less than or equal to the maximum driving parameter, the lidar is controlled to scan the two calibration boards to obtain point clouds. When the current driving parameter is greater than the maximum driving parameter, the control of the lidar to scan the two calibration boards is stopped.

[0164] For example, the driving parameter range is [-0.1, 0.1], the minimum driving parameter is -0.1, the maximum driving parameter is 0.1, the gain step value is 0.01, and the expression for the current driving parameter is: ,in, For minimum driving parameters, This is the gain step value. Let be the driving parameters for the i-th iteration. When the number of acquisitions is 1, the current driving parameters are: The lidar drives the galvanometer to rotate according to the current drive parameter -0.1, and emits a laser line through the expanding lens to scan the two calibration plates, thereby obtaining the point cloud corresponding to the current drive parameter -0.1. When the number of acquisitions is 2, the current drive parameter is... The lidar drives the galvanometer to rotate according to the current drive parameter -0.1, and emits a laser line through the expanding lens to scan the two calibration plates, thereby obtaining the point cloud corresponding to the current drive parameter -0.09. When the number of acquisitions is 12, the current drive parameter is... The lidar drives the galvanometer to rotate according to the current driving parameter 0.01, and emits a laser line through the expanding lens to scan the two calibration plates, thereby obtaining the point cloud corresponding to the current driving parameter 0.01. This process continues in the same manner.

[0165] Step S45: Determine the distance between candidate boards based on each frame of point cloud in the second point cloud set.

[0166] In this embodiment, a first feature region corresponding to the first calibration board and a second feature region corresponding to the second calibration board are extracted from each frame of point cloud in the second point cloud set. The distance between the first feature region and the second feature region is calculated and used as the distance between candidate boards.

[0167] Step S46: Determine the target driving parameters based on the distance between candidate boards and the distance between target boards.

[0168] The target driving parameters are used to drive the galvanometer to scan the target field of view through the expanding lens. In this embodiment, the absolute value of the difference between the distance between candidate plates and the distance between target plates is calculated. If the absolute value of the difference is less than or equal to a preset threshold, the driving parameter corresponding to the distance between candidate plates is determined as the target driving parameter. If the absolute value of the difference is greater than the preset threshold, the driving parameter is discarded. The preset threshold is customized by the designer based on engineering experience; for example, the preset threshold is 0.01m. This embodiment, by comparing the absolute value of the difference with the preset threshold, can find the target driving parameter with the smallest error. This ensures that when the lidar operates according to the target driving parameters, the output field of view is consistent with or has a very small error compared to the target field of view, thereby achieving precise correction of the field of view.

[0169] It can be understood that "the galvanometer scans the target field of view using the expanding lens" can be interpreted as the galvanometer scanning the target field of view using the expanding lens with a field of view that is very close to the target field of view, or it can be interpreted as the galvanometer scanning the target field of view using the expanding lens with a field of view that is consistent with the target field of view.

[0170] In summary, the embodiments of this application have at least the following technical effects:

[0171] 1) In this embodiment, the point cloud of the first point cloud set is subjected to angular distortion correction processing. The obtained target point cloud suppresses the point cloud distortion caused by the angular lens, which is conducive to obtaining a more reliable and accurate target plate distance. The target plate distance is used as the quantization target for adjusting the driving parameters. The driving parameters are continuously adjusted to obtain target driving parameters that make the candidate plate distance consistent with the target plate distance. Based on the target driving parameters, the lidar can drive the galvanometer to scan through the angular lens with the target field of view. Therefore, this embodiment can help the lidar to accurately and reliably output the laser line corresponding to the target field of view.

[0172] 2) This application embodiment does not use direct measurement to correct the field of view range, but rather collects point clouds and corrects the field of view range based on the collected point clouds. Related technologies usually require a great many measurement operations to minimize the error correction of the field of view range, while the number of data collection operations in this application embodiment is far less than the number of measurements in related technologies, enabling fast and accurate correction of the field of view range.

[0173] 3) Since the wide-angle lens is a non-linear device, it is difficult for related technologies to find the driving parameter with the smallest error to correct the error of the field of view to the ideal minimum value. However, the embodiments of this application can find the target driving parameter with the smallest error and correct the error of the field of view to the ideal minimum value, for example, the minimum value is 0.01. Therefore, the correction accuracy of the approach of the embodiments of this application is very high.

[0174] 4) Most related technologies correct the field of view without an extended-angle lens, while the embodiments of this application perform correction based on a lidar equipped with an extended-angle lens, and the correction result is very accurate.

[0175] It should be noted that in the above embodiments, there is no necessarily a certain order between the steps. Those skilled in the art can understand from the description of the embodiments of this application that the above steps may have different execution orders in different embodiments, that is, they may be executed in parallel or in turn, etc.

[0176] As another aspect of the embodiments of this application, this application provides a field-of-view range correction device. The field-of-view range correction device can be a software module, which includes several instructions stored in a memory. A processor can access the memory, call the instructions, and execute them to complete the field-of-view range correction methods described in the various embodiments above.

[0177] In some embodiments, the field-of-view correction device can also be constructed from hardware devices. For example, the field-of-view correction device can be constructed from one or more chips, which can work in coordination to complete the field-of-view correction method described in the various embodiments above. As another example, the field-of-view correction device can also be constructed from various logic devices, such as general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), microcontrollers, ARM (Acorn RISC Machine) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination of these components.

[0178] Please see Figure 8 The field-of-view correction device 800 includes: a first point cloud acquisition module 81, an angular distortion correction module 82, a first distance determination module 83, a second point cloud acquisition module 84, a second distance determination module 85, and a driving parameter acquisition module 86.

[0179] The first point cloud acquisition module 81 is used to acquire a first point cloud set, which includes multiple frames of point clouds obtained by the lidar using the same driving parameters to drive the galvanometer to scan two calibration plates spaced apart. The angular distortion correction module 82 is used to perform angular distortion correction processing on the point clouds in the first point cloud set to obtain the target point cloud. The first distance determination module 83 is used to determine the distance between the target plates based on the target point cloud. The second point cloud acquisition module 84 is used to acquire a second point cloud set, which includes multiple frames of point clouds obtained by the lidar using different driving parameters to drive the galvanometer to scan two calibration plates. The second distance determination module 85 is used to determine the distance between candidate plates based on each frame of point cloud in the second point cloud set. The driving parameter acquisition module 86 is used to determine the target driving parameters based on the distance between candidate plates and the target distance, wherein the target driving parameters are used to drive the galvanometer to scan within the target field of view range through the angular expansion lens.

[0180] In some embodiments, the expansion angle distortion correction module 82 is specifically used to: generate an expansion angle calibration curve based on a first point cloud set, wherein the expansion angle calibration curve is used to correct the point cloud distortion generated by the expansion angle lens, and correct the point cloud in the first point cloud set based on the expansion angle calibration curve to obtain the target point cloud.

[0181] In some embodiments, the first distance determination module 83 is specifically used to: determine an adjustment ratio based on the target point cloud, the adjustment ratio being used to reflect the difference between the current field of view of the lidar and the specified target field of view; determine the reference plate distance based on the first point cloud set, wherein the reference plate distance is the plate distance after the angular distortion correction processing; and determine the target plate distance based on the adjustment ratio and the reference distance.

[0182] In some embodiments, the first distance determination module 83 is further specifically used to: determine the current field of view of the lidar based on the target point cloud, and determine an adjustment ratio based on the preset target field of view and the current field of view.

[0183] In some embodiments, the target point cloud includes multiple rows of sub-point clouds, and the first distance determination module 83 is further specifically used to: select sub-point clouds that meet preset position conditions from the target point cloud as target sub-point clouds, and determine the current field of view range of the lidar based on the target sub-point clouds.

[0184] In some embodiments, the first distance determination module 83 is further specifically used to: determine the middle row index of the target point cloud, determine that the sub-point cloud corresponding to the middle row index satisfies the preset position conditions, and set the sub-point cloud corresponding to the middle row index as the target sub-point cloud.

[0185] In some embodiments, the target sub-point cloud includes multiple point cloud points, and the first distance determination module 83 is further specifically used for: acquiring the coordinate information of each point cloud point, determining the yaw angle corresponding to the point cloud point based on the coordinate information of the point cloud point, searching for the minimum yaw angle and the maximum yaw angle among the yaw angles of multiple point cloud points, and determining the current field of view of the lidar based on the minimum yaw angle and the maximum yaw angle.

[0186] In some embodiments, the first distance determination module 83 is further specifically used to: select a point cloud corresponding to the target attitude as a reference point cloud in the first point cloud set, wherein the laser line emitted by the lidar from the direction corresponding to the middle field of view under the target attitude is located in the middle of the two calibration plates, obtain the expansion angle calibration curve, correct the reference point cloud based on the expansion angle calibration curve, obtain the corrected reference point cloud, and determine the distance between the reference plates based on the corrected reference point cloud.

[0187] In some embodiments, the driving parameter acquisition module 86 is specifically used to: calculate the absolute value of the difference between the candidate board distance and the target board distance, and determine the driving parameter corresponding to the candidate board distance as the target driving parameter in response to the absolute value of the difference being less than or equal to a preset threshold.

[0188] It should be noted that the above-mentioned field-of-view range correction device can execute the field-of-view range correction method provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects of the method. Technical details not described in detail in the embodiments of the field-of-view range correction device can be found in the field-of-view range correction method provided in the embodiments of this application.

[0189] See Figure 9 , Figure 9 This is a schematic diagram of a computer device provided in an embodiment of this application. The computer device 900 includes one or more processors 91 and a memory 92. The memory 92 is connected to one or more processors 91, for example, via a bus.

[0190] Processor 91 is configured to support the computer device in performing the corresponding functions in the methods described in the above method embodiments. The processor may be a central processing unit (CPU), a network processor (NP), a hardware chip, or any combination thereof. The aforementioned hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The aforementioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0191] Memory 92 is used to store program code, etc. Memory may include volatile memory (VM), such as random access memory (RAM); memory may also include non-volatile memory (NVM), such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); memory may also include combinations of the above types of memory.

[0192] The memory 92 can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the field-of-view range correction method in the embodiments of this application. The processor executes the non-volatile software programs, instructions, and modules stored in the memory to perform various functional applications and data processing of the field-of-view range correction method and the field-of-view range correction device, thereby realizing the functions of each module or unit of the field-of-view range correction method and the field-of-view range correction device provided in the above method embodiments.

[0193] The memory 92 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and application programs required for at least one function. The data storage area may store data created based on the use of the field-of-view correction device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the field-of-view correction device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0194] The one or more modules are stored in the memory. When executed by the one or more processors, they perform the field of view range correction method in any of the above method embodiments. For example, they perform the method steps described in the above method embodiments to realize the functions of the modules described in the above device embodiments.

[0195] This application also provides a computer-readable storage medium storing a computer program, the computer program including program instructions, which, when executed by a computer device, cause the computer device to perform the method described in the foregoing embodiments.

[0196] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0197] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

Claims

1. A method for correcting the field of view, applied to lidar, characterized in that, The lidar includes a galvanometer and an expanding lens, the galvanometer and the expanding lens being arranged opposite to each other, and the correction method includes: Obtain a first point cloud set, wherein the first point cloud set includes multiple frames of point clouds obtained by the lidar using the same driving parameters to drive the galvanometer to scan two calibration plates set at intervals. Perform angular distortion correction processing on the point clouds in the first point cloud set to obtain the target point cloud; Determining the target inter-plate distance based on the target point cloud includes: determining an adjustment ratio based on the target point cloud, the adjustment ratio being used to reflect the difference between the current field of view of the lidar and the specified target field of view; determining a reference inter-plate distance based on the first point cloud set, wherein the reference inter-plate distance is an inter-plate distance after angular distortion correction processing; and determining the target inter-plate distance based on the adjustment ratio and the reference inter-plate distance. Obtain a second point cloud set, wherein the second point cloud set includes multiple frames of point clouds obtained by the lidar using different driving parameters to drive the galvanometer to scan the two calibration plates; The distance between candidate boards is determined based on each frame of point cloud in the second point cloud set; The target driving parameters are determined based on the distance between the candidate plates and the distance between the target plates, wherein the target driving parameters are used to drive the galvanometer to scan through the expanding lens within the target field of view.

2. The correction method according to claim 1, characterized in that, The step of performing angular distortion correction processing on the point cloud in the first point cloud set to obtain the target point cloud includes: An expansion angle calibration curve is generated based on the first point cloud set, wherein the expansion angle calibration curve is used to correct the point cloud distortion generated by the expansion angle lens; The target point cloud is obtained by correcting the point cloud in the first point cloud set based on the expansion angle calibration curve.

3. The correction method according to claim 1, characterized in that, The step of determining the adjustment ratio based on the target point cloud includes: The current field of view of the lidar is determined based on the target point cloud; The adjustment ratio is determined based on the preset target field of view range and the current field of view range.

4. The correction method according to claim 3, characterized in that, The target point cloud includes multiple rows of sub-point clouds, and determining the current field of view of the lidar based on the target point cloud includes: Select sub-point clouds that meet preset position conditions from the target point cloud as target sub-point clouds; The current field of view of the lidar is determined based on the target sub-point cloud.

5. The correction method according to claim 4, characterized in that, The step of selecting sub-point clouds that meet preset position conditions from the target point cloud includes: Determine the middle row index of the target point cloud; Determine that the sub-point cloud corresponding to the intermediate row index satisfies the preset position condition; Set the sub-point cloud corresponding to the intermediate row index as the target sub-point cloud.

6. The correction method according to claim 4, characterized in that, The target sub-point cloud includes multiple point cloud points, and determining the current field of view range of the lidar based on the target sub-point cloud includes: Obtain the coordinate information of each point in the point cloud; Determine the yaw angle corresponding to the point cloud point based on the coordinate information of the point cloud point; The minimum and maximum yaw angles are searched among the yaw angles of multiple point cloud points; The current field of view of the lidar is determined based on the minimum yaw angle and the maximum yaw angle.

7. The correction method according to claim 1, characterized in that, Determining the distance between reference boards based on the first point cloud set includes: In the first set of point clouds, a point cloud corresponding to the target attitude is selected as a reference point cloud, wherein the laser line emitted by the lidar from the direction corresponding to the middle field of view under the target attitude is located between the two calibration plates. Obtain the expansion angle calibration curve; The reference point cloud is corrected based on the expansion angle calibration curve to obtain the corrected reference point cloud; The distance between reference boards is determined based on the corrected reference point cloud.

8. The correction method according to any one of claims 1 to 7, characterized in that, The determination of target driving parameters based on the candidate inter-board distance and the target inter-board distance includes: Calculate the absolute value of the difference between the distance between the candidate boards and the distance between the target boards; In response to the absolute value of the difference being less than or equal to a preset threshold, the driving parameter corresponding to the distance between the candidate boards is determined as the target driving parameter.

9. A computer device, characterized in that, The device includes a memory and a processor, the memory being connected to the processor, the processor being configured to execute one or more computer programs stored in the memory, the processor causing the computer device to implement the field of view range correction method as described in any one of claims 1-8 when executing the one or more computer programs.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the field-of-view range correction method as described in any one of claims 1-8.

Images