A linear laser gear measurement sensor space pose calibration device and method

Abstract

The application discloses a linear laser gear measurement sensor space pose calibration device and method, which integrates a calibration cylindrical reference section, a calibration D-shaped reference section and a gear positioning installation section on the same core rod at the same time, does not need to calibrate and replace a gear to be measured in advance, and is consistent with a measurement reference. Therefore, reference error introduced by replacement of an installation component is avoided, and time cost of repeated operation is reduced. Meanwhile, geometric constraints are carried out on the basis of plane parameters of the calibration D-shaped reference section, axis parameters of the calibration cylindrical reference section and relative position parameters of the gear to be measured and the calibration cylindrical reference section. Therefore, space pose parameters of the linear laser measurement device relative to a gear coordinate system of the gear to be measured can be quickly and stably solved, and a problem of freedom degeneration caused by an unobservable rotation angle around an axis of the core rod in single cylindrical geometric fitting is solved.

Description

Technical Field

[0001] This application relates to the field of gear measurement technology, and in particular to a spatial pose calibration device and method for a line laser gear measurement sensor. Background Technology

[0002] For small module gears, existing contact measurement methods have certain drawbacks, such as difficulty in inserting the probe into the tooth groove, tooth deformation caused by force measurement, and low efficiency of point-by-point scanning. These issues make it difficult to meet the high-throughput inspection and online quality control requirements of the small module gear industry. Non-contact measurement is becoming an important direction for small module gear inspection, among which laser triangulation using line lasers has achieved a good overall balance between accuracy, speed, and engineering feasibility. Current line laser gear measurement solutions suffer from problems such as inconsistencies between the calibration and measurement references (calibration is performed by first installing calibration parts and then installing a mandrel to fix the gear for data acquisition and measurement), leading to errors and increased time costs. Furthermore, existing calibration techniques are not simple to operate (requiring multi-position acquisition), single geometric fitting (such as fitting only a cylinder) results in degree-of-freedom degradation, and some calibration parts are complex to manufacture. Therefore, there is an urgent need to propose a calibration scheme that integrates calibration and measurement simultaneously. Summary of the Invention

[0003] This application aims to propose a spatial pose calibration device and method for a line laser gear measurement sensor, which can realize simultaneous calibration and measurement, thereby achieving rapid and stable solution of the sensor's spatial pose, and overcoming the problem of degree of freedom degradation caused by the unobservable rotation angle around the axis when only a single cylindrical geometry is used for fitting.

[0004] In a first aspect, embodiments of this application provide a spatial pose calibration device for a line laser gear measurement sensor, comprising: Gear measurement and mounting platform; The mandrel includes, in sequence, an installation and positioning cylindrical section, a calibration cylindrical reference section, a gear positioning cylindrical section, and a calibration D-type reference section. The installation and positioning cylindrical section is fixedly connected to the gear measurement and installation platform. The gear positioning cylindrical section is used to mount the gear to be measured. The calibration D-type reference section is formed by cutting the cylindrical surface axially to form a D-shaped cross-section, and the plane reference plane of the D-shaped cross-section is parallel to the axis of the mandrel. A line laser measuring device is installed on the gear measuring mounting platform and located above the mandrel. The line laser measuring device is used to acquire the point cloud data of the cylindrical reference surface of the calibration cylindrical reference segment and the point cloud data of the planar reference surface of the calibration D-type reference segment in a single linear scan acquisition. The control and calculation module is electrically connected to the line laser measuring device. This module is used to acquire point cloud data of the cylindrical reference surface, point cloud data of the planar reference surface, and relative position parameters of the gear under test and the calibration cylindrical reference segment; to fit the planar parameters of the calibration D-type reference segment based on the planar reference surface point cloud data; to fit the axial parameters of the calibration cylindrical reference segment based on the cylindrical reference surface point cloud data; and to solve for the spatial pose parameters of the line laser measuring device relative to the gear under test in the gear coordinate system based on the planar parameters, the axial parameters, and the relative position parameters.

[0005] Secondly, embodiments of this application provide a method for spatial pose calibration of a line laser gear measuring sensor, applied to the spatial pose calibration device for a line laser gear measuring sensor as described in the first aspect embodiment above. The method for spatial pose calibration of the line laser gear measuring sensor includes: Acquire the point cloud data of the cylindrical reference surface, the point cloud data of the planar reference surface, and the relative position parameters of the gear under test and the calibrated cylindrical reference segment; The planar parameters of the calibrated D-type reference segment are obtained by fitting the point cloud data of the planar reference surface. The axial parameters of the calibrated cylindrical reference segment are obtained by fitting the point cloud data of the cylindrical reference surface. The spatial pose parameters of the line laser measuring device relative to the gear coordinate system of the gear under test are determined based on the plane parameters, the axis parameters, and the relative position parameters.

[0006] The spatial pose calibration device and method for the line laser gear measuring sensor in this application integrates a calibration cylindrical reference segment, a calibration D-type reference segment, and a gear positioning and mounting segment on the same mandrel. This eliminates the need to calibrate first and then replace the gear under test, ensuring that the calibration reference and the measurement reference are consistent. This avoids the reference error introduced by replacing mounting components and reduces the time cost of repetitive operations. Furthermore, by combining the planar parameters of the calibration D-type reference segment, the axial parameters of the calibration cylindrical reference segment, and the relative position parameters of the gear under test and the calibration cylindrical reference segment for geometric constraints, it is possible to quickly and stably solve the spatial pose parameters of the line laser measuring device relative to the gear under test in the gear coordinate system. It also solves the problem of degree-of-freedom degradation caused by the unobservable rotation angle around the mandrel axis when using single cylindrical geometric fitting.

[0007] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing this application. Attached Figure Description

[0008] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a schematic diagram of the structure of a core rod according to an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a line laser measuring device according to an embodiment of this application; Figure 3 This is a flowchart of a method for spatial pose calibration of a line laser gear measuring sensor according to an embodiment of this application; Figure 4 This is point cloud data of the gear and mandrel to be tested according to an embodiment of this application; Figure 5 This is a schematic diagram of planar reference surface point cloud data and cylindrical reference surface point cloud data according to an embodiment of this application; Figure 6 This is a schematic diagram of the plane normal vector of the target plane according to an embodiment of this application; Figure 7 This is a schematic diagram of the axis vector and center point of one embodiment of this application; Figure 8 This is a schematic diagram of the initial axis direction and the orthogonal basis perpendicular to the initial axis direction according to an embodiment of this application; Figure 9 This is a schematic diagram of the point cloud in the mandrel coordinate system according to an embodiment of this application; Figure 10 This is a point cloud diagram in the gear coordinate system according to an embodiment of this application.

[0009] Figure label: Installation positioning cylindrical section 110, calibration cylindrical reference section 120, gear positioning cylindrical section 130, calibration D-type reference section 140; Gear under test: 200; Linear slide 310, line laser sensor 320. Detailed Implementation

[0010] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0011] In the description of this application, the use of terms such as "first," "second," etc., is for the purpose of distinguishing technical features only and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated or the order of the technical features indicated.

[0012] In the description of this application, it should be understood that the orientation descriptions, such as up, down, etc., are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0013] In the description of this application, it should be noted that, unless otherwise explicitly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0014] The technical solution of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are some embodiments of this application, not all embodiments.

[0015] See Figure 1 and Figure 2 , Figure 1 This is a schematic diagram of the structure of a mandrel according to an embodiment of this application. Figure 2 This is a schematic diagram of the structure of a line laser measuring device according to an embodiment of this application.

[0016] This application provides a spatial pose calibration device for a line laser gear measurement sensor, including a gear measurement mounting platform, a mandrel, a line laser measurement device, and a control and calculation module.

[0017] The mandrel includes, in sequence, a mounting and positioning cylindrical section 110, a calibration cylindrical reference section 120, a gear positioning cylindrical section 130, and a calibration D-type reference section 140. The mounting and positioning cylindrical section 110 is fixedly connected to the gear measuring and mounting platform. The gear positioning cylindrical section 130 is used to mount the gear 200 to be measured. The calibration D-type reference section 140 is formed by cutting the cylindrical surface axially to form a D-shaped cross section, and the plane reference plane of the D-shaped cross section is parallel to the axis of the mandrel. A line laser measuring device is installed on the gear measuring mounting platform and located above the mandrel. The line laser measuring device is used to acquire point cloud data of the cylindrical reference surface of the calibration cylindrical reference segment 120 and point cloud data of the planar reference surface of the calibration D-type reference segment 140 in a single linear scan acquisition. The control and calculation module is electrically connected to the line laser measuring device. This module is used to acquire point cloud data of the cylindrical reference surface, point cloud data of the planar reference surface, and the relative position parameters of the gear under test 200 and the calibration cylindrical reference segment 120; to obtain the planar parameters of the calibration D-type reference segment 140 based on the planar reference surface point cloud data; to obtain the axial parameters of the calibration cylindrical reference segment 120 based on the cylindrical reference surface point cloud data; and to solve for the spatial pose parameters of the line laser measuring device relative to the gear under test 200 in the gear coordinate system based on the planar parameters, axial parameters, and relative position parameters.

[0018] In this embodiment, by simultaneously integrating the calibration cylindrical reference section 120, the calibration D-type reference section 140, and the gear positioning mounting section on the same mandrel, it is not necessary to calibrate first and then replace the gear 200 under test. The calibration reference and the measurement reference are consistent, which avoids the reference error introduced by replacing the mounting components and reduces the time cost of repetitive operations. At the same time, by combining the planar parameters of the calibration D-type reference section 140, the axial parameters of the calibration cylindrical reference section 120, and the relative position parameters of the gear 200 under test and the calibration cylindrical reference section 120 for geometric constraints, it is possible to quickly and stably solve the spatial pose parameters of the linear laser measuring device relative to the gear 200 under test in the gear coordinate system. It also solves the problem of degree of freedom degradation caused by the unobservable rotation angle around the mandrel axis when performing geometric fitting of a single cylinder.

[0019] The aforementioned calibration cylindrical reference section 120 is used to provide a cylindrical reference surface to determine the axial direction reference of the line laser measuring device relative to the mandrel.

[0020] The aforementioned D-type reference segment 140 is used to provide a directional reference for breaking rotational symmetry.

[0021] The mandrel has certain precision requirements: the coaxiality of each cylindrical section should meet 1µm; the cylindricity of the mounting positioning cylindrical section 110 and the gear positioning cylindrical section 130 should meet 1µm; the calibration cylindrical reference section 120 should meet a total runout error of less than 1µm relative to the axis of the gear positioning cylindrical section 130; and the plane of the calibration D-type reference section 140 should meet a flatness of 1µm and a parallelism of less than 1µm between itself and the axis of the calibration cylindrical reference section 120.

[0022] The dimensions of the mandrel are known and used for subsequent joint nonlinear optimization of planar reference surface point cloud data and cylindrical reference surface point cloud data; the radius of the calibrated cylindrical reference segment 120 is... The distance from a point on the axis of the calibrated cylindrical reference section 120 to the plane reference surface of the calibrated D-type reference section 140 is... The axial distance between the center of the gear mounting position and the center of the calibrated cylindrical reference section 120 is... (i.e., the relative position parameters between the gear under test 200 and the calibration cylindrical reference section 120).

[0023] In some embodiments, the line laser measuring device includes a linear slide 310 and a line laser sensor 320.

[0024] A linear slide 310 is mounted on a gear measuring and mounting platform and located above the mandrel. A linear slide rail is located below the linear slide 310. The line laser sensor 320 is slidably connected to the linear slide rail via a driving component. The driving component is used to drive the line laser sensor 320 to slide on the linear slide rail, so that the line laser sensor 320 can acquire the point cloud data of the cylindrical reference surface of the calibration cylindrical reference segment 120 and the point cloud data of the planar reference surface of the calibration D-type reference segment 140 in one linear scan acquisition.

[0025] During operation, the linear slide 310 drives the linear laser sensor 320 along a specified direction. The device moves and outputs a movement signal to the line laser sensor 320. The line laser sensor 320 simultaneously scans the target object (mandrel and gear 200 under test), sequentially scanning the calibration D-type reference section 140 of the mandrel, the gear positioning cylindrical section 130, the gear 200 under test, the calibration cylindrical reference section 120, and the installation positioning cylindrical section 110.

[0026] The scanning direction proceeds sequentially from the calibration D-type reference section 140 of the mandrel to the installation and positioning cylindrical section 110. The mandrel coordinate system is constructed as follows: the axis of the cylindrical reference surface of the calibration cylindrical reference section 120 is used as the Y-axis of the mandrel coordinate system, and the normal vector of the plane reference surface of the calibration D-type reference section 140 is used as the Z-axis of the mandrel coordinate system; gear coordinate system. With mandrel coordinate system The Z-axis and X-axis are parallel but not collinear, while the Y-axis is collinear and in the same direction; and the linear laser coordinate system The three attitude angles relative to the mandrel coordinate system are limited to a certain range (greater than -90° and less than 90°), and the rotation order is fixed as ZYX, in order to avoid gimbal lock and to ensure that the subsequent rotation matrix decouples to the only Euler angle in engineering.

[0027] The formula for rigid transformation is: ;Formula (1) in, Let be the coordinates of the point in the mandrel coordinate system. , Let the coordinates of the point be in the online laser coordinate system. , It is a rotation matrix consisting of three Euler angles (rotation angles about the X-axis). Rotation angle around the Y-axis Rotation angle around the Z-axis )constitute. The constraint formula is: ; Formula (2) Meanwhile, the three Euler angles can be solved by the following formula: ;Formula (3) It is a translation vector, consisting of three translation parameters. constitute, , Let X be the translation in the X direction. This represents the translation in the Y direction. This represents the translation in the Z direction.

[0028] The three Euler angles (attitude parameters) and three translation parameters (position parameters) mentioned above constitute the spatial pose of the line laser sensor 320 relative to the gear coordinate system.

[0029] In this embodiment, the spatial pose parameters are solved in two stages. First, the plane normal vector of the plane reference surface is solved using the point cloud data of the plane reference surface, and the axis vector and center point are solved using the point cloud data of the cylindrical reference surface. Then, the X-axis direction of the mandrel coordinate system is solved by the cross product of the plane normal vector and the axis vector, and an orthogonal matrix is ​​constructed to represent the initial rotation matrix. The initial translation vector is solved using the center point and the initial rotation matrix. Furthermore, a joint optimization method is used to solve the spatial pose parameters as a whole to improve the robustness of the solution. Finally, the spatial pose parameters of the final linear laser sensor 320 are solved based on the relative position parameters of the gear under test 200 and the calibration cylindrical reference segment 120.

[0030] The reason for adopting joint optimization in this application is that the planar reference surfaces for calibrating the cylindrical reference segment 120 and the D-type reference segment 140 are obtained by point cloud fitting. Affected by factors such as point cloud noise, outliers, occlusion, and uneven point density, the two types of fitting results are usually inconsistent in actual data and cannot strictly satisfy ideal geometric relationships (such as perfectly orthogonal / parallel directions, perfectly consistent positions, etc.). If the initial pose value is directly constructed based on the single fitting result, it is easy to cause the accumulation of coupling errors between attitude and translation, which in turn affects the stability and repeatability of the calibration results. In particular, the planar reference surface for calibrating the D-type reference segment 140 is used to overcome the problem of unobservable orientation around the axis caused by the rotational symmetry of the cylinder. Joint optimization makes this orientation constraint work together with the cylinder axis constraint, avoiding the degree of freedom degradation that occurs when calibrating with only a single geometry.

[0031] See Figure 3 , Figure 3This is a flowchart of a spatial pose calibration method for a line laser gear measuring sensor according to an embodiment of this application. This method is applied to the spatial pose calibration device for a line laser gear measuring sensor described in the first aspect embodiment above. The method includes steps S100 to S400: Step S100: Obtain point cloud data of cylindrical datum surface, point cloud data of planar datum surface, and relative position parameters of the gear under test 200 and the calibration cylindrical datum segment 120; Step S200: Based on the point cloud data of the plane reference surface, the plane parameters of the calibration D-type reference segment 140 are obtained by fitting. Step S300: Obtain the axis parameters of the calibrated cylindrical reference segment 120 by fitting the point cloud data of the cylindrical reference surface; Step S400: Solve the spatial pose parameters of the line laser measuring device relative to the gear coordinate system of the gear under test 200 based on the plane parameters, axis parameters, and relative position parameters.

[0032] In this embodiment, by simultaneously integrating the calibration cylindrical reference section 120, the calibration D-type reference section 140, and the gear positioning mounting section on the same mandrel, it is not necessary to calibrate first and then replace the gear 200 under test. The calibration reference and the measurement reference are consistent, which avoids the reference error introduced by replacing the mounting components and reduces the time cost of repetitive operations. At the same time, by combining the planar parameters of the calibration D-type reference section 140, the axial parameters of the calibration cylindrical reference section 120, and the relative position parameters of the gear 200 under test and the calibration cylindrical reference section 120 for geometric constraints, it is possible to quickly and stably solve the spatial pose parameters of the linear laser measuring device relative to the gear 200 under test in the gear coordinate system. It also solves the problem of degree of freedom degradation caused by the unobservable rotation angle around the mandrel axis when performing geometric fitting of a single cylinder.

[0033] See Figure 4 , Figure 4 This is point cloud data of the gear 200 and mandrel under test according to an embodiment of this application. In step S100 above, point cloud data including the gear 200 and mandrel under test are obtained by scanning. Next, the mandrel point cloud was processed separately: Since the scanning direction was known, the mandrel point cloud was initially separated based on prior information, i.e., the y-coordinate of the point cloud; then, the density-based noise spatial clustering algorithm DBSCAN was used to remove outliers, and clustering was performed to obtain cylindrical reference surface point cloud data, planar reference surface point cloud data, and other outlier point cloud data; the two classes with the largest number of points were selected as target point clouds (because the calibration cylindrical reference segment 120 and the calibration D-type reference segment 140 occupy the main volume of the mandrel scanning area, and the corresponding number of point clouds is much larger than other interference points), and the coordinates of the center points of the two classes of point clouds were calculated, and the one with the smaller y-value was selected as the planar reference surface point cloud data of the calibration D-type reference segment 140. The point cloud data of the cylindrical reference surface with a larger y-value is used as the calibration cylindrical reference segment 120. The scanned point clouds are all in the online laser coordinate system.

[0034] It should be noted that, in addition to basing the analysis on the y-coordinate value of the center point, the two types of point clouds can also be processed separately using the RANSAC algorithm to obtain the inlier ratio. The point cloud data with the larger inlier ratio is set as the planar reference surface point cloud data. Conversely, smaller values ​​are set as point cloud data on a cylindrical reference surface. .

[0035] The principle of the RANSAC algorithm is existing technology known to those skilled in the art, and will not be elaborated here.

[0036] See Figure 5 , Figure 5 This is a schematic diagram of planar reference surface point cloud data and cylindrical reference surface point cloud data according to an embodiment of this application. After obtaining the planar reference surface point cloud data and cylindrical reference surface point cloud data, outlier removal can be performed to remove isolated noise points and local anomalies; further smoothing processing can be performed to reduce the impact of local measurement noise on the geometric fitting results.

[0037] See Figure 6 , Figure 6 This is a schematic diagram of the plane normal vector of the target plane according to an embodiment of this application. In some embodiments, step S200 specifically includes, but is not limited to, steps S210 to S230: Step S210, from the point cloud data of the plane reference surface Three points are randomly selected from the middle to construct an initial plane, resulting in multiple initial planes; Step S220: Calculate the point cloud data of the planar reference surface respectively. For each point, the point-to-plane distance from each initial plane is calculated. Points whose point-to-plane distance is less than a preset distance threshold are recorded as interior points. The initial plane with the most interior points is taken as the target plane. Step S230: Calculate the plane normal vector of the target plane. The plane normal vector of the target plane As a plane parameter.

[0038] The formula for the general equation of a plane is: ;Formula (4) in, Let be the three-dimensional unit normal vector of the plane to be fitted, used to characterize the spatial orientation and geometric pose of the plane. It is a three-dimensional point coordinate vector in space. The plane position constant term represents the normal directional distance from the fitted plane to the origin of the coordinate system.

[0039] The initial solution for the target plane is achieved by iterating the RANSAC algorithm for a predetermined number of times (10-20 times). In each iteration, the point with the farthest distance from the plane is removed, with a removal ratio of 2%-5%. After multiple iterations, the initial plane with the most interior points is selected as the target plane.

[0040] See Figure 7 , Figure 7 This is a schematic diagram of the axis vector and center point according to an embodiment of this application. In some embodiments, step S300 specifically includes, but is not limited to, steps S310 to S370: Step S310: Perform principal component analysis on the point cloud data of the cylindrical reference surface to obtain the direction of the first principal component; Step S320: The direction of the first principal component is taken as the initial axis direction of the calibration cylindrical reference segment 120; Step S330: Construct a cross-sectional plane with the initial axis direction as the normal, and project each point in the cylindrical reference surface point cloud data onto the cross-sectional plane to obtain multiple two-dimensional coordinate points; Step S340: Establish an elliptical boundary model based on multiple two-dimensional coordinate points, and calculate the radial residual of each point in the cylindrical reference surface point cloud data based on the elliptical boundary model. Step S350: Construct an objective function based on the radial mean square of each radial residual, optimize the objective function with the goal of minimizing the residual, and robustly remove the point cloud data of the cylindrical datum surface based on the radial residual after each round of parameter update to obtain the axis vector of the calibrated cylindrical datum segment 120 and the point cloud of the target cylindrical datum surface. Step S360: Project the centroid of the point cloud of the target cylindrical reference surface onto the axis vector to obtain the center point; Step S370: Use the axis vector and center point as axis parameters.

[0041] In some implementations, step S310 specifically includes, but is not limited to, steps S311 to S314: Step S311: Decentralize and construct the covariance matrix for the cylindrical reference surface point cloud data; Step S312: Perform eigenvalue decomposition on the covariance matrix to obtain multiple eigenvalues ​​and corresponding eigenvectors sorted by size. Step S313: Denote the eigenvector corresponding to the largest eigenvalue as the first principal component eigenvector; Step S314: Take the direction of the first principal component feature vector as the first principal component direction.

[0042] In steps S310 and S320, it can be understood that the first principal component eigenvector corresponding to the largest eigenvalue reflects the direction of the largest spatial distribution variance of the cylindrical reference surface point cloud data. This direction is consistent with the axis of the cylinder itself. Therefore, the first principal component eigenvector is used as the initial axis direction for cylinder fitting.

[0043] See Figure 8 , Figure 8 This is a schematic diagram of the initial axial direction and the orthogonal basis perpendicular to the initial axial direction according to an embodiment of this application. In step S330, the cross-sectional plane with the initial axial direction as the normal is constructed through the following steps: First, construct a set of orthogonal bases perpendicular to the initial axis direction. The method for constructing an orthogonal basis is as follows: Select a direction that is perpendicular to the initial axis. A non-parallel reference vector , It can be , If so, then the orthogonal bases are respectively , .

[0044] Downsampling is used to reduce the number of points in the cylindrical reference surface point cloud data, thereby reducing the computational load. Then, the centroid of the cylindrical reference surface point cloud data is calculated using the following formula: ;Formula (5) in, The centroid of the point cloud data for the cylindrical reference surface. For points in the point cloud data of the cylindrical reference surface, This represents the number of point clouds in the cylindrical reference surface point cloud data.

[0045] For any point in the point cloud data of the cylindrical reference surface Calculate its relative centroid. (or a vector at a point on the axis) Project its component along the initial axis direction as And obtain its component perpendicular to the axis. ,Will On orthogonal basis Expanding upwards, we obtain the two-dimensional coordinates of the cross-sectional plane. ,in, , Project all points in the cylindrical reference surface point cloud data onto the initial axis direction. In the cross-sectional plane with the normal direction, a two-dimensional coordinate point set is obtained. The polar angle and polar radius of the cross-sectional points are calculated. Then, the radial distance from the point in the cylindrical datum surface point cloud data to the initial axis is calculated. for: ;Formula (6) And define its polar angle for: ;Formula (7) In steps S340 and S350, considering the possibility of slight ellipticization or measurement errors in the cylindrical reference surface, the cross-sectional profile is modeled as an ellipse, which is oriented as... Theoretical radius for: ;Formula (8) in, Let be the lengths of the two semi-axes of the elliptical cross section.

[0046] Constructing radial residuals And the mean square of the radial residuals of all sample points is used as the objective function: ;Formula (9) in, This is a point on the initial axis (the reference point for the cylinder's center line). This is the unit vector in the direction of the initial axis.

[0047] Constrained by bounded optimization within a given radius. The following solution yields the updated result. , The radius tolerance of the preset calibration cylindrical reference section 120 can be 0.005mm.

[0048] To suppress the influence of outliers, side angle transition regions, and reflective noise on the axis solution, robust removal of the cylindrical reference surface point cloud data is performed based on the radial residual after each round of parameter updates. The preferred method is based on the median absolute deviation (...). Thresholding strategy: Let the median of the radial residual set of the current point cloud be... , for Then the threshold for removal is... Take as M AD ,in, As an empirical coefficient, it can be set to 3. Retain those that satisfy the condition. The points are used as the new set of inliers, and the maximum elimination ratio in each round is limited to a preset value, while ensuring that the proportion of inliers is not lower than a preset lower limit, so as to avoid unstable fitting caused by over-elimination. Repeat the outer loop iteration of "parameter optimization - residual elimination" until the radial root mean square error change between two adjacent iterations is less than a threshold or the set of interior points no longer changes, finally obtaining the axis direction vector of the cylindrical reference plane, i.e., the axis vector. And the new target cylindrical reference surface point cloud after removing points. .

[0049] In steps S360 and S370, the centroid is further projected onto the axis vector to obtain the center point. It can be used to define the origin of the target coordinate system or to solve for translation.

[0050] In some implementations, step S400 specifically includes, but is not limited to, steps S410 to S450: Step S410: Construct the initial rotation matrix and initial translation vector based on the plane normal vector, axis vector, and center point; Step S420: Based on the point cloud contained in the target plane Point cloud of the target cylindrical reference surface The target rotation matrix is ​​obtained by performing global nonlinear optimization on the initial rotation matrix; Step S430: Construct an intermediate translation matrix based on the relative position parameters; Step S440: Form the target translation matrix based on the intermediate translation matrix and the initial translation vector; Step S450: The target rotation matrix and target translation matrix are used as the spatial pose parameters of the line laser measuring device relative to the gear coordinate system of the gear under test 200.

[0051] In some implementations, step S410 specifically includes, but is not limited to, steps S411 to S415: Step S411: Normalize the axis vector to obtain the Y-axis of the mandrel coordinate system; Step S412: Orthogonalize the plane normal vector to obtain the Z-axis of the mandrel coordinate system; Step S413: Perform the cross product of the axis vector and the plane normal vector to obtain the X-axis of the mandrel coordinate system; Step S414: Construct an orthogonal matrix based on the X-axis, Y-axis and Z-axis of the mandrel coordinate system to obtain the initial rotation matrix; Step S415: Solve for the initial translation vector based on the center point and the initial rotation matrix.

[0052] To ensure the consistency of direction in the subsequent establishment of the coordinate system, the fitted axis vectors are first... With plane normal vector Apply sign constraints to orient it toward a predetermined positive direction: axis vector Should point to If pointing to Then take the opposite. Plane normal vector Should point to If pointing to Then take the opposite. .

[0053] An orthogonal coordinate basis is constructed using the axis vector and the plane normal vector. The mandrel coordinate system is defined to satisfy the following conditions: the Y-axis of the mandrel coordinate system is aligned with the axis of the cylinder, and the Z-axis of the mandrel coordinate system is aligned with the plane normal vector and orthogonal to the Y-axis.

[0054] In step S411 above, the normalized axis vector is used as the Y-axis of the mandrel coordinate system. : ;Formula (10) in, Represents the mandrel coordinate system Directional representation of axes in the online laser coordinate system.

[0055] In step S412 above, the plane normal vector is orthogonalized to obtain the Z-axis of the mandrel coordinate system. Because fitting error may lead to and Not strictly orthogonal, Gram-Schmidt orthogonalization is used to remove its in... Components in direction: ;Formula (11) ;Formula (12) in, Represents the mandrel coordinate system Directional representation of axes in the online laser coordinate system To remove in Components in direction in the mandrel coordinate system Directional representation of axes in the online laser coordinate system.

[0056] In step S413 above, the X-axis of the core coordinate system is determined by the cross product. And ensure that the right hand is tied: ;Formula (13) in, Represents the mandrel coordinate system Directional representation of axes in the online laser coordinate system.

[0057] Corrected again by cross product Axis to reduce numerical error: .

[0058] In step S414 above, the direction cosine matrix is ​​constructed and the initial rotation matrix is ​​obtained. Arrange the representations of the mandrel coordinate system in the three-axis online laser coordinate system into a matrix by column: .because It is an orthogonal matrix that satisfies Therefore, the initial rotation matrix from the line laser coordinate system to the mandrel coordinate system... for: .

[0059] In step S415 above, the reference point on the axis is mapped to the origin of the mandrel coordinate system. To ensure that the selected axis reference point has a zero vector in the mandrel coordinate system, let... ,in, Using the center point, we obtain the initial translation vector. : .

[0060] The initial pose is now obtained. ,formula Used to transform point clouds in line laser coordinates to mandrel coordinates.

[0061] The reason for using Gram-Schmidt is that the cylinder axis and the plane normal obtained by actual fitting are often not strictly orthogonal (or have slight deviations) under noise, and the rotation matrix must be composed of three strictly orthogonal unit axes. Gram-Schmidt is equivalent to "projecting another direction onto its orthogonal complement space and normalizing it while keeping the main constraint direction as unchanged as possible", thereby constructing a set of strictly orthogonal, right-handed, unit coordinate bases, ensuring that the initial rotation matrix obtained is a legal orthogonal rotation matrix, and also providing a more stable initial value for subsequent nonlinear optimization.

[0062] In some implementations, step S420 specifically includes, but is not limited to, steps S421 to S424: Step S421: Transform the point cloud contained in the target plane from the line laser coordinate system to the mandrel coordinate system according to the initial rotation matrix to obtain the mandrel plane point cloud; Step S422: Transform the point cloud of the target cylindrical reference surface from the line laser coordinate system to the mandrel coordinate system according to the initial rotation matrix to obtain the point cloud of the mandrel cylindrical reference surface. Step S423: Construct a planar residual term based on the mandrel planar point cloud, construct a cylindrical residual term based on the mandrel cylindrical reference surface point cloud, and concatenate the planar residual term and the cylindrical residual term into a unified residual vector; Step S424: Establish a robust nonlinear least squares objective based on the unified residual vector and solve it to obtain the objective rotation matrix.

[0063] This application uses interior points instead of all points for joint optimization, which reduces the impact of outliers on the solution. Residual weights are set to balance the two types of constraints. Since the number of points in the two point sets may differ, a weighting coefficient is set to prevent the class with more points from dominating the objective function. and In some implementations, the weight of the cylindrical term is set to... Planar term weights Proportional to the square root of the ratio of points: ,in, The number of points in the point cloud contained in the target plane. The number of points in the cloud of points on the target cylindrical reference surface.

[0064] A pose parameterization form is established, representing the pose of the line laser sensor as a combination of rotation and translation vectors, i.e. , where the rotation vector Translation vector , These are the three components of the Rodrigues rotation vector. The vector direction represents the direction of the equivalent rotation axis, and the magnitude represents the rotation angle around that axis. The Rodrigues formula can be used to convert between the rotation vector and the rotation matrix, and it is used to describe the attitude of the line laser sensor coordinate system relative to the reference coordinate system. These are the three-dimensional coordinate components of the origin of the line laser sensor coordinate system in the reference coordinate system, used to describe the position of the line laser sensor coordinate system relative to the reference coordinate system.

[0065] Rotation vector Using the initial rotation matrix The calculation shows that the translation vector is the same as the initial translation vector, and the initial parameters are obtained by concatenating the two. , where the rotation vector Calculated from the antisymmetric terms of the matrix: ;Formula (14) Among them, rotation angle Calculated using the trace formula: ;Formula (15) Let the rotation vector be... Its modulus is the rotation angle: ;when At that time, the unit axis of rotation is: Construct antisymmetric matrix : ;Formula (16) in, Rotate the X-axis by units. Rotate the Y-axis by units. Rotate the Z-axis in units of 1.

[0066] Then the rotation matrix can be obtained. : ;Formula (17) in, It is an identity matrix. It should be noted that when... At that time, it can be made To avoid numerical instability.

[0067] In steps S421 and S422, for any candidate parameter Calculate the rotation matrix The point cloud of the target cylindrical reference surface and the point cloud contained in the target plane are transformed from the linear laser coordinate system to the mandrel coordinate system: ;Formula (18) ;Formula (19) in, Let represent the point cloud coordinates of the target cylindrical reference surface in the line laser coordinate system. Let the coordinates of the target cylindrical reference surface be the point cloud coordinates in the mandrel coordinate system. equal , Let be the point cloud coordinates contained in the target plane in the line laser coordinate system. The coordinates of the point cloud contained in the target plane in the mandrel coordinate system are given. This yields the point cloud of the mandrel plane and the point cloud of the mandrel cylindrical reference surface.

[0068] In step S423, the ideal cylindrical axis is defined in the mandrel coordinate system as... The axis passes through the origin of the mandrel coordinate system, and the design radius is... Then the cylindrical point radial distance for: ;Formula (20) Cylinder residuals Defined as: .

[0069] In the mandrel coordinate system, the ideal plane is defined as: , , Let Z be the length component of a point on an ideal plane in the mandrel coordinate system along the Z-axis. To determine the distance from a point on the axis of the cylindrical reference segment 120 to the plane reference surface for calibrating the type D reference segment 140, the plane residual term... Defined as: .

[0070] The cylindrical residual term and planar residuals Concatenate into a unified residual vector : ;Formula (21) in, The weights of the cylindrical residual terms. The weights of the planar residual terms.

[0071] In step S424, a nonlinear least squares optimization problem is constructed: ; Formula (22) because It is a nonlinear function, at the current iteration value Perform a first-order Taylor expansion on the uniform residual vector: ;Formula (23) ;Formula (24) in, For Jacobian matrices, For increments.

[0072] Substituting formulas (23) and (24) into formula (22), we obtain a quadratic approximation. : ;Formula (25) in, , Expanding and ignoring higher-order terms, we can obtain information about... The quadratic form: ;Formula (26) Again To find the extreme value, we obtain the normal equation of the Gauss-Newton iterative algorithm (GN algorithm): ;Formula (27) To enhance numerical stability and convergence region, a damping term can be introduced using the LM algorithm: ;Formula (28) in, The damping factor, It is the identity matrix. When the value is large, it approximates gradient descent; when the value is small, it approximates the normal equation of the GN algorithm.

[0073] Solving for the results Then, update the parameters. Iterate until any of the following stopping conditions are met: (1) the descent amount of formula (22) (2) Parameter increment norm (3) The maximum number of iterations is reached.

[0074] Finally, the optimal parameters are obtained. The rotation parameters are obtained through Rodrigues' formula. Convert to target rotation matrix That is, recovery and , It is equal to the initial translation vector.

[0075] In some implementations, the target rotation matrix and the initial translation vector are substituted back into the point cloud contained in the target plane. Point cloud of the target cylindrical reference surface In this process, the RMS / maximum value of the cylindrical residual term and the root mean square / maximum value of the planar residual term are calculated as calibration quality evaluation indicators and used to determine whether the calibration is valid.

[0076] It should be noted that the principles of Gram-Schmidt orthogonalization, Rodriguez formula, GN algorithm, and LM algorithm mentioned above are existing technologies known to those skilled in the art, and will not be elaborated here.

[0077] The embodiment of this application adopts separate global joint optimization because the preceding planar fitting and cylindrical fitting are performed separately and are affected by noise outliers and local occlusion. The initial pose obtained by directly combining the results of the two is only a usable initial value of "geometric consistency approximation" and cannot guarantee that it is optimal at the same time under the same objective function. Joint optimization puts the cylindrical constraint and the planar constraint into the same nonlinear least squares problem to minimize them together. It can redistribute the error globally, correct the deviation caused by fitting and orthogonalization, and obtain the spatial pose parameters that are most consistent with the two types of references at the same time. This improves accuracy and stability and reduces sensitivity to initial values ​​and local outliers.

[0078] The aforementioned rotation parameters can also be represented using Euler angles or quaternions, which are equivalent to the rotation vector form. This application's embodiments preferably use rotation vectors to ensure that the rotation matrix satisfies orthogonality constraints and is numerically stable during iteration, making it more suitable for continuous optimization, while also avoiding the gimbaling phenomenon of Euler angles.

[0079] See Figure 9 and Figure 10, Figure 9 This is a schematic diagram of the point cloud in the mandrel coordinate system according to an embodiment of this application. Figure 10 This is a point cloud diagram in the gear coordinate system according to an embodiment of this application.

[0080] In step S430, considering that the gear coordinate system is fixed to the left side of the mandrel coordinate system, close to the negative half-axis of the Y-axis of the gear coordinate system, an intermediate translation matrix is ​​constructed. .

[0081] In step S440, the target translation matrix for ,in, It is equal to the initial translation vector.

[0082] In step S450, the target rotation matrix and target translation matrix are used as the spatial pose parameters of the line laser measuring device relative to the gear coordinate system of the gear under test 200. The three Euler angles, i.e., the attitude parameters, can be obtained by decoupling the final target rotation matrix; the three translation parameters, i.e., the position parameters, can be obtained by decoupling the final target translation matrix.

[0083] The control and computing module in this application embodiment can be an electronic device or a component within an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile perSonal computer (UMPC), netbook, or perSonal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), perSonal computer (PC), television, ATM, or self-service machine, etc. This application embodiment does not specifically limit the scope of the electronic device.

[0084] This application also provides an electronic device, including a processor and a memory. The memory stores a program or instructions that can run on the processor. When the program or instructions are executed by the processor, they implement the various steps of the above-described embodiment of the spatial pose calibration method for the line laser gear measuring sensor and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0085] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described embodiment of the spatial pose calibration method for the linear laser gear measuring sensor and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0086] It should be clarified that this application is not limited to the specific configurations and processes described above. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and illustrated as examples. However, the method process of this application is not limited to the specific steps described and illustrated. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0087] The functional blocks described above can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. The programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0088] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0089] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A linear laser gear measurement sensor space pose calibration device, characterized in that, include: Gear measurement and mounting platform; The mandrel includes, in sequence, an installation and positioning cylindrical section, a calibration cylindrical reference section, a gear positioning cylindrical section, and a calibration D-type reference section. The installation and positioning cylindrical section is fixedly connected to the gear measurement and installation platform. The gear positioning cylindrical section is used to mount the gear to be measured. The calibration D-type reference section is formed by cutting the cylindrical surface axially to form a D-shaped cross-section, and the plane reference plane of the D-shaped cross-section is parallel to the axis of the mandrel. A line laser measuring device is installed on the gear measuring mounting platform and located above the mandrel. The line laser measuring device is used to acquire the point cloud data of the cylindrical reference surface of the calibration cylindrical reference segment and the point cloud data of the planar reference surface of the calibration D-type reference segment in a single linear scan acquisition. The control and calculation module is electrically connected to the line laser measuring device. This module is used to acquire point cloud data of the cylindrical reference surface, point cloud data of the planar reference surface, and relative position parameters of the gear under test and the calibration cylindrical reference segment; to fit the planar parameters of the calibration D-type reference segment based on the planar reference surface point cloud data; to fit the axial parameters of the calibration cylindrical reference segment based on the cylindrical reference surface point cloud data; and to solve for the spatial pose parameters of the line laser measuring device relative to the gear under test in the gear coordinate system based on the planar parameters, the axial parameters, and the relative position parameters.

2. The linear laser gear measurement sensor space pose calibration apparatus of claim 1, wherein, The line laser measurement device includes: A linear slide is disposed on the gear measuring and mounting platform and located above the mandrel, and a linear slide rail is provided below the linear slide; A line laser sensor is slidably connected to the linear slide rail via a driving component. The driving component is used to drive the line laser sensor to slide on the linear slide rail, so that the line laser sensor can acquire the point cloud data of the cylindrical reference surface of the calibration cylindrical reference segment and the point cloud data of the planar reference surface of the calibration D-type reference segment in one linear scan acquisition.

3. A line laser gear measurement sensor space pose calibration method, characterized in that, The linear laser gear measuring sensor spatial pose calibration device as described in claim 1 or 2, wherein the linear laser gear measuring sensor spatial pose calibration method comprises: Acquire the point cloud data of the cylindrical reference surface, the point cloud data of the planar reference surface, and the relative position parameters of the gear under test and the calibrated cylindrical reference segment; The planar parameters of the calibrated D-type reference segment are obtained by fitting the point cloud data of the planar reference surface. The axial parameters of the calibrated cylindrical reference segment are obtained by fitting the point cloud data of the cylindrical reference surface. The spatial pose parameters of the line laser measuring device relative to the gear coordinate system of the gear under test are determined based on the plane parameters, the axis parameters, and the relative position parameters.

4. The line laser gear measurement sensor space pose calibration method of claim 3, wherein, The process of fitting the plane parameters of the calibrated D-type reference segment based on the point cloud data of the plane reference surface includes: Three points are randomly selected from the point cloud data of the plane reference surface to construct an initial plane, resulting in multiple initial planes; Calculate the point-to-plane distance from all points in the planar reference surface point cloud data to each initial plane. Points whose point-to-plane distance is less than a preset distance threshold are recorded as interior points. The initial plane with the most interior points is taken as the target plane. Calculate the plane normal vector of the target plane, and use the plane normal vector of the target plane as the plane parameter.

5. The line laser gear measurement sensor space pose calibration method of claim 4, wherein, The process of fitting the axial parameters of the calibrated cylindrical reference segment based on the point cloud data of the cylindrical reference surface includes: Principal component analysis was performed on the point cloud data of the cylindrical reference surface to obtain the direction of the first principal component; The direction of the first principal component is taken as the initial axis direction of the calibration cylindrical reference segment; A cross-sectional plane with the initial axis direction as the normal is constructed, and each point in the cylindrical reference surface point cloud data is projected onto the cross-sectional plane to obtain multiple two-dimensional coordinate points; An elliptical boundary model is established based on multiple two-dimensional coordinate points, and the radial residual of each point in the cylindrical reference surface point cloud data is calculated based on the elliptical boundary model. An objective function is constructed based on the radial mean square of each radial residual. The objective function is optimized and solved with the goal of minimizing the residual. After each round of parameter update, the point cloud data of the cylindrical reference surface is robustly removed based on the radial residual to obtain the axis vector of the calibrated cylindrical reference segment and the point cloud of the target cylindrical reference surface. The centroid of the target cylindrical reference surface point cloud is projected onto the axis vector to obtain the center point; The axis vector and the center point are used as the axis parameters.

6. The line laser gear measurement sensor space pose calibration method of claim 5, wherein, The step of performing principal component analysis on the point cloud data of the cylindrical reference surface to obtain the first principal component direction includes: The covariance matrix is ​​constructed by decentralizing the point cloud data of the cylindrical reference surface; The covariance matrix is ​​decomposed into eigenvalues ​​to obtain multiple eigenvalues ​​and corresponding eigenvectors sorted by size. The eigenvector corresponding to the largest eigenvalue is denoted as the first principal component eigenvector. The direction of the first principal component feature vector is taken as the direction of the first principal component.

7. The line laser gear measurement sensor space pose calibration method of claim 5, wherein, The step of solving the spatial pose parameters of the line laser measuring device relative to the gear coordinate system of the gear under test based on the plane parameters, the axis parameters, and the relative position parameters includes: Construct an initial rotation matrix and an initial translation vector based on the plane normal vector, the axis vector, and the center point; The initial rotation matrix is ​​optimized nonlinearly based on the point cloud contained in the target plane and the point cloud of the target cylindrical reference surface to obtain the target rotation matrix; Construct an intermediate translation matrix based on the relative position parameters; The target translation matrix is ​​formed based on the intermediate translation matrix and the initial translation vector; The target rotation matrix and the target translation matrix are used as the spatial pose parameters of the line laser measuring device relative to the gear coordinate system of the gear under test.

8. The line laser gear measurement sensor space pose calibration method of claim 7, wherein, The step of constructing the initial rotation matrix and initial translation vector based on the plane normal vector, the axis vector, and the center point includes: Normalize the axis vector to obtain the Y-axis of the mandrel coordinate system; Orthogonalize the plane normal vector to obtain the Z-axis of the mandrel coordinate system; The cross product of the axis vector and the plane normal vector is used to obtain the X-axis of the mandrel coordinate system; The initial rotation matrix is ​​obtained by constructing an orthogonal matrix based on the X, Y, and Z axes of the mandrel coordinate system. The initial translation vector is obtained by solving based on the center point and the initial rotation matrix.

9. The line laser gear measurement sensor space pose calibration method of claim 8, wherein, The constraint formula for the initial translation vector is: ； wherein, is the initial rotation matrix, is the center point, is the initial translation vector.

10. The line laser gear measurement sensor space pose calibration method of claim 7, wherein, The step of performing overall nonlinear optimization of the initial rotation matrix based on the point cloud contained in the target plane and the point cloud of the target cylindrical reference surface to obtain the target rotation matrix includes: The point cloud contained in the target plane is transformed from the line laser coordinate system to the mandrel coordinate system according to the initial rotation matrix to obtain the mandrel plane point cloud; The point cloud of the target cylindrical reference surface is transformed from the line laser coordinate system to the mandrel coordinate system according to the initial rotation matrix to obtain the point cloud of the mandrel cylindrical reference surface. A planar residual term is constructed based on the planar point cloud of the mandrel, and a cylindrical residual term is constructed based on the cylindrical reference surface point cloud of the mandrel. The planar residual term and the cylindrical residual term are then concatenated into a unified residual vector. A robust nonlinear least squares objective is established and solved based on the unified residual vector to obtain the objective rotation matrix.

Images