Multi-line laser sensor global calibration method based on calibration block

Abstract

The invention relates to the technical field of line laser sensor detection and calibration, in particular to a multi-line laser sensor global calibration method based on a calibration block, which is characterized by comprising the following steps: independently designing the calibration block: designing a hexagonal prism-shaped calibration block, the overlook surface of the calibration block is provided with a hexagon of A-B-C-D-E-F, the angle E and the angle F are right angles, a line segment BC is parallel to a line segment EF, and the line segment BC is parallel to the line segment EF; wherein the lengths of the inclined planes AB and CD are 5 mm, the standard tolerance is + / -0.002 mm, the angle D is 120 degrees, and the angle A is 150 degrees; global calibration is realized through establishment of a world coordinate system, calibration error analysis, feature interval extraction and coarse registration. The method has the effects of high efficiency and high precision in the calibration process.

Description

Technical Field

[0001] The present invention relates to the technical field of sensor global calibration, and in particular to a multi-line laser sensor global calibration method based on a calibration block. Background Art

[0002] The prior application, numbered 202411562409.2, discloses a global calibration method for multiple line laser sensors, belonging to the field of laser sensor detection technology. This method addresses the technical challenge of designing a specialized, high-precision three-dimensional target for simultaneous global calibration of multiple line laser sensors. The key technical solution involves using different line laser sensors to acquire contour data from different directions of a three-dimensional calibration block, extracting the common field of view data from each sensor based on curvature features. The common field of view of each sensor is then coarsely and finely aligned, and the transformation matrix of each sensor is then determined to achieve global calibration of the multi-sensor system.

[0003] According to the specifications of the line laser sensor, a calibration block needs to be designed separately. Without a calibration block, multiple line laser sensors cannot work together. If they all use their own independent coordinate systems, global calibration cannot be completed. If each sensor is calibrated separately, it will be very time-consuming and inefficient. Summary of the Invention

[0004] In order to solve the technical problems and shortcomings in the prior art, the present invention provides a global calibration method for a multi-line laser sensor based on a calibration block, which can overcome the technical problems of low efficiency and accuracy in improving the global calibration of line laser sensors.

[0005] To achieve the above-mentioned and other related purposes, the present invention adopts the following technical solutions: A global calibration method for a multi-line laser sensor based on a calibration block, comprising: Calibration block design and production: Design a hexagonal prism-shaped calibration block. The top view of the calibration block has a hexagonal shape of ABCDEF, where ∠E and ∠F are right angles, line segment BC is parallel to line segment EF, the length of the two inclined surfaces AB and CD is 5mm, the standard tolerance is ±0.002mm, ∠D is 120°, and ∠A is 150°; Establishment of the world coordinate system: By matching the common field of view of different sensors, the transformation matrix from each sensor coordinate system to the world coordinate system is calculated to achieve global calibration of multiple sensors; Calibration error analysis: By adjusting the translation stage where the sensor is located, the sensor can be accurately projected onto the calibration block, and the laser sensors on each line are made as parallel as possible, and the deflection angle error is evaluated; Feature interval extraction: Based on the estimated deflection angle error, during the calibration process, we first need to ensure that the light planes of each sensor are as coplanar as possible. Then, we extract the contour images of the calibration blocks captured by each sensor, find the common field of view between the sensors in the image coordinate system, and accurately align the common fields of view of different sensors. Based on the local curvature, we obtain the coordinates of the feature points in the common fields of view of different sensors. Coarse registration: After obtaining the coordinates of the feature points in each sensor coordinate system, the coordinate system of sensor one is used as the reference coordinate system. Based on the principle of rigid transformation, the feature point vectors of different coordinate systems are registered and the coordinate system transformation matrix is ​​solved to lay the foundation for subsequent fine registration.

[0006] Preferably, it also includes fine alignment: the fine alignment steps are as follows: (1) Using bidirectional KD tree from target point cloud Internal search source point cloud midpoint The nearest neighbor of , while in the source point cloud Search target point cloud middle The nearest neighbor of ; until and For the same point, and Form a one-to-one matching point pair, (2) Calculate the rotation matrix between corresponding point pairs and translation matrices , and calculate the mean square error of the transformation matrix , making Minimum, where: ; (3) Determine the iteration termination condition and set the threshold , and set the maximum number of iterations, when Or the iteration is terminated when the maximum number of iterations is reached, otherwise the calculated transformation matrix transforms the source point cloud and updates the point cloud , and repeat steps (1) to (3), and the final transformation matrix can be obtained after iteration and ,The coarsely registered point cloud is precisely registered on the target point cloud.

[0007] Preferably, the method for estimating the measurement error of the deflection angle includes: estimating using the following formula: ; ;in is the deflection angle, is the actual contour value of the object, is the profile value measured by the line laser sensor, is the error between the measured contour and the actual contour.

[0008] Compared with the prior art, the present invention has the following beneficial effects: 1. To achieve high-precision and high-efficiency requirements in multi-sensor global calibration, the present invention designs and processes a special high-precision three-dimensional calibration block according to the specifications of the line laser sensor. The calibration block design can directly or indirectly create a common field of view and feature points in the measurement area of ​​each line laser sensor, facilitating global calibration of each sensor.

[0009] 2. In the present invention, by matching the common fields of view of different sensors and calculating the transformation matrix from the coordinate system of each sensor to the world coordinate system, global calibration of multiple sensors can be achieved with very high efficiency and accuracy.

[0010] Other additional advantages and benefits of the present application will be given in part in the following description, and in part will become apparent from the following description, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is a schematic diagram of the calibration block structure of an embodiment of the present application; Figure 2 This is a schematic diagram of multi-sensor synchronous calibration according to an embodiment of the present application; Figure 3 This is a schematic diagram of a sensor coordinate system according to an embodiment of the present application; Figure 4 is a schematic diagram of the deflection angle calibration principle of an embodiment; Figure 5 is a profile image captured by each sensor in the embodiment. DETAILED DESCRIPTION

[0012] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0013] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. The illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component can be changed at will, and the component layout may also be more complex.

[0014] It should be noted that, in the description of the present application, the terms "up", "down", "left", "right", "inside", "outside" and the like indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, which are merely for the convenience of description, and do not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the invention. In addition, it should also be noted that, in the description of the present application, unless otherwise clearly specified and limited, the terms "install", "connect", "connect" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the two elements. For those skilled in the art, the specific meanings of the above terms in the invention can be understood according to the specific circumstances. Example

[0015] The present invention discloses a global calibration method for a multi-line laser sensor based on a calibration block.

[0016] like Figure 1 As shown, first design the calibration block: In order to achieve high precision and high efficiency in multi-sensor global calibration, a special high-precision three-dimensional calibration block was designed and processed according to the specifications of the line laser sensor. Figure 1 and Figure 2 The length of the two inclined surfaces AB and CD is 5mm, with a standard tolerance of ±0.002mm. This calibration block design allows each line laser sensor's measurement area to directly or indirectly have a common field of view and feature points, facilitating global calibration of each sensor.

[0017] The second step is to establish the coordinate system: First, a world coordinate system is established based on the pose of the sensor's measurement range, such as Figure 3 As shown. Assume that the measurement coordinate system of sensor 1 is , let the origin of the world coordinate system be With the origin of the measurement coordinate system Overlap, let and Establish a world coordinate system by measuring the U and V directions of the coordinate system respectively . Multiple sensors are being calibrated simultaneously 、 、 and are the image coordinate systems of sensors one, two, three, and four, respectively. A, B, C, and D are feature points on the calibration block with known distances. A and B are within the common field of view of sensors one, two, and three, and the corresponding image points in the image coordinate system of sensor one are and , in the image coordinate system of sensor 2 and , in the image coordinate system of sensor three and , C and D are within the common field of view of sensors 2, 3, and 4, and the corresponding image points in the image coordinate system of sensor 2 are and , in the image coordinate system of sensor three and , in the image coordinate system of sensor 4 and By matching the common field of view of different sensors and finding the transformation matrix from each sensor coordinate system to the world coordinate system, global calibration of multiple sensors can be achieved. Define the pose of the measurement range of sensor 1 to establish the world coordinate system, and transform the coordinate systems of other sensors to the world coordinate system, that is, to the image coordinate system of sensor 1. The transformation matrix from the coordinate system of sensor i to the coordinate system of sensor 1 is , and its transformation formula is as follows:

[0018] in is the rotation matrix for transforming each sensor coordinate system to the sensor 1 coordinate system, is the displacement vector converted from each sensor coordinate system to the sensor 1 coordinate system.

[0019] Perform calibration error analysis: Before calibration, the laser planes of each sensor need to be aligned as much as possible. We adjust the displacement stage where the sensor is located so that the line laser can be accurately projected onto the calibration block and each line laser is as parallel as possible. However, due to certain errors in the actual installation and adjustment process, it is difficult for the laser planes of each sensor to be completely aligned, resulting in a deflection angle between the actual plane and the theoretical plane, such as Figure 4 This deflection angle introduces measurement error, which affects measurement accuracy. To assess the impact of this error, the following formula can be used for estimation.

[0020] ; ; in is the deflection angle, is the actual contour value of the object, is the profile value measured by the line laser sensor, is the error between the measured contour and the actual contour.

[0021] About feature interval extraction: Based on the above error sources, during the calibration process, we first need to ensure that the light planes of each sensor are as coplanar as possible, and then extract the contour images of the calibration blocks taken by each sensor, such as Figure 3-5 By finding the common fields of view between sensors in the image coordinate system, such as fields AB and CD, and accurately aligning the common fields of view of different sensors, the calibration process is finally completed.

[0022] The contour data of the line laser projected onto the cross section of the calibration block is obtained. The curvature at the edge of the contour reaches its maximum value in its neighborhood. Therefore, the coordinates of the feature points in the common field of view of different sensors can be obtained based on the local curvature. The curvature of the two-dimensional contour can be solved using the three-point circle method, as shown below: ; Where a, b and c are three adjacent points of the point cloud contour 、 and The length of the side of the triangle is , and S is the area of ​​the triangle. Due to the influence of noise during the measurement process, some point cloud data is prone to floating. The position and height of the curvature peak are affected by noise, causing the location of feature points to deviate from their actual physical positions, making it impossible to accurately locate all feature points using the local extremum method. To improve calibration accuracy and accurately extract feature point coordinates, the Savitzky-Golay algorithm is used to filter the contour data. This algorithm applies the least squares method within each moving window and performs a polynomial fit on the data points to obtain a smoothed contour. Edge feature points are found by calculating the curvature of each point on the contour curve.

[0023] Regarding coarse registration: After obtaining the coordinates of the feature points in each sensor coordinate system, in order to achieve spatial unification of multi-view data, it is necessary to use the coordinate system of sensor one as the reference coordinate system, align the feature point vectors of different coordinate systems based on the principle of rigid transformation, and solve the coordinate system transformation matrix to lay the foundation for subsequent fine registration.

[0024] Assume that the feature point vectors of the reference coordinate system and the coordinate system to be registered are , . Normalized eigenvector: ; ; Calculate the rotation angle of the coordinate system to be registered relative to the reference coordinate system: . are unit vectors Relative to The rotation angle of the coordinate system to be registered relative to the reference coordinate system can be obtained. : ; Calculation of translation matrix for coarse registration: .

[0025] About fine registration: Although the two-dimensional rigid transformation method can achieve coarse registration of two sets of contours, it can only handle two linear transformations, rotation and translation, and has limited ability to handle nonlinear errors, and cannot correct more complex geometric mismatches. To improve matching accuracy, the bidirectional iterative closest point (DCP) algorithm is used for fine registration. By establishing a bidirectional KD tree (K-dimensiontree) for bidirectional search, compared with the traditional unidirectional iterative closest point (ICP) algorithm, it can more accurately capture matching point pairs, effectively avoid the problem of local optimal solutions, and significantly improve the registration accuracy of complex contour data.

[0026] Defined as contour data after rough registration Source point cloud, base coordinate system contour data is the target point cloud, where and is the total amount of precisely registered point cloud data, and The steps for fine registration are as follows: (1) Using bidirectional KD tree from target point cloud Internal search source point cloud midpoint The nearest neighbor of , while in the source point cloud Search target point cloud middle The nearest neighbor of ; until and For the same point, and Form a one-to-one matching point pair, (2) Calculate the rotation matrix between corresponding point pairs and translation matrices , and calculate the mean square error of the transformation matrix , making Minimum, where: ; (3) Determine the iteration termination condition and set the threshold , and set the maximum number of iterations, when Or the iteration is terminated when the maximum number of iterations is reached, otherwise the calculated transformation matrix transforms the source point cloud and updates the point cloud , and repeat steps (1) to (3), and the final transformation matrix can be obtained after iteration and ,The coarsely registered point cloud is precisely registered on the target point cloud.

[0027] By aligning the contours of the calibration blocks under each sensor, global calibration of multiple sensors can be achieved, and the final rotation matrix is ​​shown as follows: , the displacement vector is shown as follows: .

[0028] The technical solutions of the present invention have been described so far in conjunction with the preferred embodiments shown in the accompanying drawings. The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A global calibration method for a multi-line laser sensor based on a calibration block, characterized in that: include, Calibration block design and production: Design a hexagonal prism-shaped calibration block. The top view of the calibration block has a hexagonal shape of ABCDEF, where ∠E and ∠F are right angles, line segment BC is parallel to line segment EF, the length of the two inclined surfaces AB and CD is 5mm, the standard tolerance is ±0.002mm, ∠D is 120°, and ∠A is 150°; Establishment of the world coordinate system: By matching the common field of view of different sensors, the transformation matrix from each sensor coordinate system to the world coordinate system is calculated to achieve global calibration of multiple sensors; Calibration error analysis: By adjusting the translation stage where the sensor is located, the sensor can be accurately projected onto the calibration block, and the laser sensors on each line are made as parallel as possible, and the deflection angle error is evaluated; Feature interval extraction: Based on the estimated deflection angle error, during the calibration process, we first need to ensure that the light planes of each sensor are as coplanar as possible. Then, we extract the contour images of the calibration blocks captured by each sensor, find the common field of view between the sensors in the image coordinate system, and accurately align the common fields of view of different sensors. Based on the local curvature, we obtain the coordinates of the feature points in the common fields of view of different sensors. Coarse registration: After obtaining the coordinates of the feature points in each sensor coordinate system, the coordinate system of sensor one is used as the reference coordinate system. Based on the principle of rigid transformation, the feature point vectors of different coordinate systems are registered and the coordinate system transformation matrix is ​​solved to lay the foundation for subsequent fine registration.

2. The global calibration method for a multi-line laser sensor based on a calibration block according to claim 1, characterized in that: It also includes fine alignment: the steps of fine alignment are as follows: (1) Using bidirectional KD tree from target point cloud Internal search source point cloud midpoint The nearest neighbor of , while in the source point cloud Search target point cloud middle The nearest neighbor of ; until and For the same point, and Form a one-to-one matching point pair, (2) Calculate the rotation matrix between corresponding point pairs and translation matrices , and calculate the mean square error of the transformation matrix , making Minimum, where: ； (3) Determine the iteration termination condition and set the threshold , and set the maximum number of iterations, when Or the iteration is terminated when the maximum number of iterations is reached, otherwise the calculated transformation matrix transforms the source point cloud and updates the point cloud , and repeat steps (1) to (3), and the final transformation matrix can be obtained after iteration and ,The coarsely registered point cloud is precisely registered on the target point cloud.

3. The global calibration method for a multi-line laser sensor based on a calibration block according to claim 1, characterized in that: The method for estimating the measurement error of the deflection angle includes: using the following formula for estimation: ; ;in is the deflection angle, is the actual contour value of the object, is the profile value measured by the line laser sensor, is the error between the measured contour and the actual contour.

Images