A global calibration method and device for external parameters of multiple line laser 3D sensors

Abstract

The present invention discloses a global calibration method for external parameters of multiple line laser 3D sensors. The method comprises: aligning a measuring laser of the 3D sensor with a calibration line to measure and obtain the 3D coordinates under the 3D sensor; measuring along the calibration line by a three-coordinate measuring arm to obtain the 3D coordinates of the calibration line under the three-coordinate measuring arm; using a straight line fitting algorithm to obtain the point cloud coordinates of the 3D coordinates under the 3D sensor and the 3D coordinates under the three-coordinate measuring arm, respectively obtaining the broken line segments of a sawtooth target; using a straight line intersection algorithm to respectively calculate the broken line segments of the sawtooth target, and obtaining the 3D coordinates of the inflection points on the calibration line; aligning the 3D coordinates of multiple inflection points to obtain a conversion matrix between the line laser 3D sensor coordinates and the three-coordinate measuring arm coordinates, thereby obtaining the external parameter conversion matrix between different line laser 3D sensor coordinate systems, and completing the global calibration.

Description

Technical Field

[0001] The invention relates to a method and a device for global calibration of external parameters of multiple line laser 3D sensors, belonging to the technical field of visual measurement. Background Art

[0002] Line laser 3D sensors are a type of non-contact sensor based on laser measurement that is currently widely used. They can directly measure the 3D coordinates of the laser light strip formed by the laser on the surface of the object being measured, and are widely used in dimensional measurement, three-dimensional scanning and other occasions. However, the measurement coordinates of each line laser 3D sensor are based on its own coordinate system. When multiple line laser 3D sensors are used, the outputs of different sensors cannot be unified in one coordinate system, and system-level measurements cannot be completed. At present, the methods for achieving multi-sensor global calibration generally include dual-target calibration and multi-target calibration, but they all require opening the visual camera interface inside the sensor to obtain the original image information, and then use the image-based visual calibration method to achieve the external parameter global calibration of multiple sensors. Secondly, it can also be achieved using the dedicated calibration tool provided by the line laser 3D sensor manufacturer itself, but this generally requires the use of sensors of the same brand, which is not applicable to sensors of different brands and types, which brings inconvenience to some applications.

[0003] At the same time, since the 3D measurement points of the line laser 3D sensor are all located on the laser plane, one dimension is actually missing. Therefore, the traditional method of using 3D targets to calibrate the positional relationship between multiple 3D sensors cannot be used directly. Summary of the invention

[0004] The purpose of the present invention is to provide a method and device for global calibration of external parameters of multiple line laser 3D sensors, so as to solve the defect that the traditional method of using 3D targets to realize position relationship calibration between multiple 3D sensors cannot be applied to different types of sensors.

[0005] A global calibration method for external parameters of multiple line laser 3D sensors, the method comprising:

[0006] Project the laser of the line laser 3D sensor onto the calibration line of the target, so that the alignment mark is on the laser light bar and the laser light bar falls on the calibration line;

[0007] The measurement laser of the 3D sensor is aligned with the calibration line to measure the 3D coordinates under the 3D sensor;

[0008] Measure along the calibration line through the three-coordinate measuring arm to obtain the 3D coordinates of the calibration line under the three-coordinate measuring arm;

[0009] The point cloud coordinates of the 3D coordinates under the 3D sensor and the 3D coordinates under the three-dimensional coordinate measuring arm are respectively obtained through a straight line fitting algorithm to obtain the broken line segments of the sawtooth target;

[0010] The line segments of the sawtooth target are calculated using the straight line intersection algorithm to obtain the 3D coordinates of the inflection points on the calibration line;

[0011] By aligning the 3D coordinates of multiple inflection points, the transformation matrix between the line laser 3D sensor coordinates and the three-coordinate measurement arm coordinates is obtained, thereby obtaining the external parameter transformation matrix between different line laser 3D sensor coordinate systems to complete the global calibration.

[0012] Furthermore, the 3D coordinate system under the 3D sensor is S -xyz, the measured point cloud is:

[0013] L S ={s1,s2,...,sn}. The set of all point clouds is represented as a segment of a polyline. Using the straight line fitting algorithm, the straight line equation of each segment of the polyline is calculated, ax+by+cz+1=0, so that:

[0014]

[0015] Let the equations of two adjacent straight line segments p and q be a p x+b p y+c p z+1=0、a q x+b q y+c q z+1=0; calculate the midpoint of the perpendicular line segment between the two straight line segments as the approximate point of the inflection point, and record all the inflection points as SS={ss 1 ,ss 2 ,...,ss t}.

[0016] Furthermore, the coordinate system of the three-coordinate measuring arm is R -xyz, the measured point cloud is: L R ={r1,r2,...,rm}, and calculate the linear equation of each segment of the polyline by ax+by+cz+1=0, so that:

[0017]

[0018] Let the equations of two adjacent straight line segments p and q be a p x+b p y+c p z+1=0、a q x+b q y+c q z+1=0, calculate the midpoint of the perpendicular line segment between the two straight line segments, which is the approximate point of the inflection point, denoted as RR={rr 1 ,rr 2,...,rr t}.

[0019] Furthermore, the inflection point SS and RR are in a one-to-one correspondence in space, and the least square method is used to solve the equation: RR = R S-R ·SS+T S-R , and obtain the coordinate system O S -xyz to O R -xyz transformation relationship, rotation matrix R S-R and the translation matrix T S-R .

[0020] Furthermore, the 3D sensor has a conversion relationship of: GG = R G-R -1 R S-R SS+R G-R -1 (T S-R -T G-R ).

[0021] Furthermore, there are three alignment marks, which are arranged at both ends and in the middle of the calibration line, and the three alignment marks are not collinear.

[0022] A global calibration device for external parameters of multiple line laser 3D sensors, comprising a sawtooth target, a three-coordinate measuring arm and a line laser 3D sensor;

[0023] The sawtooth target is provided with at least two sawteeth and a calibration line, and alignment marks are provided at both ends and in the middle of the calibration line;

[0024] The three-coordinate measuring arm is arranged at the end of the sawtooth target and is used to measure the 3D coordinates of the calibration line under the three-coordinate measuring arm;

[0025] The line laser 3D sensor is arranged above the sawtooth target. The measuring laser of the line laser 3D sensor is aligned with the calibration line through the alignment marks at both ends and in the middle of the calibration line to measure the 3D coordinates of the line laser 3D sensor.

[0026] Furthermore, the calibration line is arranged perpendicular to the saw teeth and is coplanar with the saw teeth.

[0027] Compared with the prior art, the present invention has the following beneficial effects: the present method further expands the flexibility of external parameter calibration of multiple line laser 3D sensors. First, it does not need to utilize the image information of the sensor and does not need to develop an internal interface. Second, it can be used for line laser 3D sensors of different brands and types, such as those based on the triangulated vision principle or emission type. Third, there is no requirement for the measurement field of view of multiple sensors and no requirement for a common field of view. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0029] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0030] The present invention proposes a global calibration method for multiple line laser 3D sensors external parameters, the method comprising the following steps:

[0031] The alignment marks at both ends and in the middle of the calibration line are used to align the measuring laser of the line laser 3D sensor with the calibration line, and the 3D coordinates under the sensor are measured.

[0032] Use the three-dimensional coordinate measuring arm to measure along the calibration line to obtain the 3D coordinates of the calibration line under the three-dimensional coordinate measuring arm.

[0033] The point cloud coordinates in the above two coordinate systems are respectively used to obtain the broken line segments of the sawtooth target through a straight line fitting algorithm;

[0034] Using the straight line intersection algorithm, the 3D coordinates of no less than 3 inflection points on the calibration line are obtained.

[0035] By aligning the 3D coordinates of multiple inflection points in the two coordinate systems, the transformation matrix between the line laser 3D sensor coordinate system and the three-coordinate measuring arm coordinate system is obtained.

[0036] For other line laser 3D sensors, repeat the above process to obtain the transformation matrix between the other line laser 3D sensors and the coordinate system of the three-coordinate measurement arm.

[0037] Through the three-coordinate measuring arm coordinate system as the medium, the external parameter conversion matrix between different line laser 3D sensor coordinate systems is obtained to complete the global calibration.

[0038] (1) Move the sawtooth target so that the laser light bar of a certain line laser 3D sensor (referred to as S sensor) passes through the alignment mark on the calibration line at the same time. At this time, the laser light bar should be exactly aligned with the calibration line. Fix the sawtooth target position.

[0039] (2) Read the laser light strip point cloud measurement result of the sensor, that is, the 3D coordinates of the light strip, and denote the sensor coordinate system as O S -xyz. The measured point cloud is:

[0040] L S ={s1,s2,...,sn}. The set of all point clouds is represented as a segment of a polyline. Using the straight line fitting algorithm, the straight line equation of each segment of the polyline is calculated, ax+by+cz+1=0, so that:

[0041]

[0042] Let the equations of two adjacent straight line segments p and q be a p x+b p y+c p z+1=0、a q x+b q y+c q z+1=0. Calculate the midpoint of the perpendicular line segment between the two straight line segments and consider it as the approximate point of its inflection point. All inflection points are recorded as SS={ss 1 ,ss 2 ,...,ss t}.

[0043] (3) Using the three-dimensional coordinate measuring arm, along the calibration line of the sawtooth target, record a point cloud coordinate every small step length, that is, measure the point cloud coordinate of the calibration line in the three-dimensional coordinate measuring arm coordinate system, and record the three-dimensional coordinate measuring arm coordinate system as O R -xyz. The measured point cloud is: L R ={r1,r2,...,rm}. The set of all point clouds is represented as a segment of a polyline. Using the straight line fitting algorithm, the straight line equation of each segment of the polyline is calculated, ax+by+cz+1=0, so that:

[0044]

[0045] Let the equations of two adjacent straight line segments p and q be a p x+b p y+c p z+1=0、a q x+b q y+c q z+1=0. Calculate the midpoint of the perpendicular line segment between the two straight line segments and consider it as the approximate point of its inflection point, denoted by RR={rr 1 ,rr 2 ,...,rr t}.

[0046] (1) The inflection points SS and RR have a one-to-one correspondence in space. The following equations are solved using the least squares method to obtain the coordinate system O S -xyz to O R -xyz conversion relationship, that is, the rotation matrix R S-R and the translation matrix T S-R .

[0047] RR=R S-R ·SS+T S-R

[0048] (2) For the new sensor G, keep the position of the three-dimensional coordinate measuring arm unchanged and repeat steps (1) to (4) to obtain the sensor G coordinate system to OR -xyz conversion relationship, how to get the rotation matrix R G-R and the translation matrix T G-R .

[0049] (3) Then the conversion relationship from sensor S to G is:

[0050] GG=R G-R -1 R S-R SS+R G-R -1 (T S-R -T G-R )

[0051] If there are more than two sensors, the above steps are used to complete the conversion relationship between all sensors and complete the global calibration.

[0052] The present invention also discloses a global calibration device for external parameters of multiple line laser 3D sensors, the device comprising at least two saw teeth and a calibration line, and alignment marks are set at both ends and in the middle of the calibration line;

[0053] The three-coordinate measuring arm is arranged at the end of the sawtooth target and is used to measure the 3D coordinates of the calibration line under the three-coordinate measuring arm;

[0054] The line laser 3D sensor is arranged above the sawtooth target, and the measuring laser of the line laser 3D sensor is aligned with the calibration line through the alignment marks at both ends and in the middle of the calibration line to measure the 3D coordinates of the line laser 3D sensor;

[0055] A calibration target in the form of a sawtooth and including a calibration line, which is a narrow scratch on the surface of the sawtooth target, arranged perpendicular to the sawtooth and maintained in the same plane;

[0056] Set three alignment marks at both ends and in the middle of the calibration line, and ensure that they are not collinear;

[0057] Project the laser of the line laser 3D sensor onto the calibration line of the target, ensure that the three alignment marks are on the laser light bar, and then ensure that the laser light bar falls on the calibration line. At this time, fix the target. Read the point cloud measurement coordinates of the sensor and record them.

[0058] Use the three-coordinate measuring arm to measure the coordinates of each point on the calibration line in the three-coordinate measuring arm coordinate system along the calibration line according to the appropriate step length.

[0059] Using the point cloud coordinates of the sensor, the straight line fitting algorithm and the straight line intersection algorithm, the coordinate values ​​of each inflection point on the sawtooth target calibration line (including the top and bottom points of the sawtooth) are calculated and recorded.

[0060] Using the point cloud coordinates of the three-coordinate measuring arm, the straight line fitting algorithm and the straight line intersection algorithm are used to calculate the coordinate values ​​of each inflection point (including the top and bottom points of the sawtooth) on the sawtooth target calibration line and record them.

[0061] The coordinate values ​​of each inflection point on the target calibration line under the sensor and under the three-coordinate measurement arm are in one-to-one correspondence. By solving the linear equation by the least squares method, the rotation and translation matrix (transformation matrix) of the sensor coordinate system and the three-coordinate measurement arm coordinate system can be obtained.

[0062] Using the above principle, the transformation matrix of all line laser 3D sensor coordinate systems and three-coordinate measurement arm coordinate systems can be obtained, thereby obtaining the transformation matrix between each sensor and completing the global calibration of external parameters.

[0063] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A global calibration method for extrinsic parameters of multiple line laser 3D sensors, characterized in that: The method comprises: Project the laser of the line laser 3D sensor onto the calibration line of the sawtooth target, so that the alignment mark is on the laser light bar and the laser light bar falls on the calibration line; The measurement laser of the 3D sensor is aligned with the calibration line to measure the 3D coordinates under the 3D sensor; Measure along the calibration line through the three-coordinate measuring arm to obtain the 3D coordinates of the calibration line under the three-coordinate measuring arm; The point cloud coordinates of the 3D coordinates under the 3D sensor and the 3D coordinates under the three-dimensional coordinate measuring arm are respectively obtained through a straight line fitting algorithm to obtain the broken line segments of the sawtooth target; The line segments of the sawtooth target are calculated using the straight line intersection algorithm to obtain the 3D coordinates of the inflection points on the calibration line; By aligning the 3D coordinates of multiple inflection points, the conversion matrix between the line laser 3D sensor coordinates and the three-coordinate measurement arm coordinates is obtained, thereby obtaining the external parameter conversion matrix between different line laser 3D sensor coordinate systems to complete the global calibration; The 3D coordinate system under the 3D sensor is S -xyz, the measured point cloud is: L S ={s1,s2,...,sn}, the set of all point clouds is represented as a segment of a polyline, and the linear equation of each segment of the polyline is calculated using the linear fitting algorithm, ax+by+cz+1=0, so that: Let the equations of two adjacent straight line segments p and q be a p x+b p y+c p z+1=0、a q x+b q y+c q z+1=0; calculate the midpoint of the perpendicular line segment between the two straight line segments as the approximate point of the inflection point, and record all the inflection points as SS={ss1,ss2,...,ss t }; Wherein, the sawtooth target is provided with at least two sawteeth and a calibration line, and alignment marks are provided at both ends and in the middle of the calibration line; The three-coordinate measuring arm is arranged at the end of the sawtooth target, and is used to measure the 3D coordinates of the calibration line under the three-coordinate measuring arm; the line laser 3D sensor is arranged above the sawtooth target, and the measuring laser of the line laser 3D sensor aligns the calibration line through the alignment marks at both ends and in the middle of the calibration line to measure the 3D coordinates of the line laser 3D sensor.

2. The global calibration method for extrinsic parameters of multiple line laser 3D sensors according to claim 1, characterized in that: The coordinate system of the three-coordinate measuring arm is R -xyz, the measured point cloud is: L R ={r1,r2,...,rm}, and calculate the linear equation of each segment of the polyline by ax+by+cz+1=0, so that: Let the equations of two adjacent straight line segments p and q be a p x+b p y+c p z+1=0、a q x+b q y+c q z+1=0, calculate the midpoint of the perpendicular line segment between the two straight line segments, which is the approximate point of the inflection point, denoted by RR={rr1,rr2,...,rr t }.

3. The global calibration method for extrinsic parameters of multiple line laser 3D sensors according to claim 2, characterized in that: The inflection point SS and RR are in one-to-one correspondence in space, and the least square method is used to solve the equation: RR = R S-R ·SS+T S-R , and obtain the coordinate system O S -xyz to O R -xyz transformation relationship, rotation matrix R S-R and the translation matrix T S-R .

4. The global calibration method for extrinsic parameters of multiple line laser 3D sensors according to claim 3, characterized in that: The conversion relationship of the 3D sensor is: GG = R G-R -1 R S-R SS+R G-R -1 (T S-R -T G-R ).

5. The global calibration method for extrinsic parameters of multiple line laser 3D sensors according to claim 1, characterized in that: The number of the alignment marks is 3, which are arranged at both ends and in the middle of the calibration line, and the 3 alignment marks are not collinear.

6. The global calibration method for extrinsic parameters of multiple line laser 3D sensors according to claim 1, characterized in that: The marking line is arranged perpendicular to the saw teeth and is coplanar with the saw teeth.

Images