阅读说明：本技术 用于多线激光雷达与相机联合标定的标定方法 (Calibration method for multi-line laser radar and camera combined calibration ) 是由 裴凌 郁文贤 刘海春 李岚臻 李扬 吴奇 于 2020-11-25 设计创作，主要内容包括：一种用于多线激光雷达与相机联合标定的标定方法,通过本标定板可以一次性地实现像素平面与激光点对的关联匹配,简化激光点信息与像素点信息的匹配流程。本发明通过标准化雷达标记与进行标定板模块化拆分和动态重构提升其泛用性与易用性。(A calibration method for multi-line laser radar and camera combined calibration can realize the association matching of a pixel plane and a laser point pair at one time through the calibration plate and simplify the matching process of laser point information and pixel point information. The invention improves the universality and the usability of the radar marker by standardizing the radar marker and carrying out calibration plate modularization splitting and dynamic reconstruction.)

1. A calibration method for the combined calibration of a multi-line laser radar and a camera is characterized by comprising the following steps:

step 1) moving a calibration plate bracket until the laser radar point is coincident with the plane of the Aruco pixel, adjusting the posture and leveling;

step 2) adsorbing the camera marking standard plate on a calibration plate support, calibrating internal parameters of the camera, inputting a calibration result into a calibration module, adjusting the pose of the camera marking standard plate, adjusting the front-back distance of the calibration plate support, and locking the frame after a proper position is found;

step 3) opening laser radar visualization software, adjusting the upper and lower positions of the laser radar marking standard plate according to the reflection intensity, and placing the laser radar marking standard plate;

step 4) reading corresponding height value h of each mark standard plate_c,h_l1,h_l2,h_l3,...；

Step 5) converting points on the radar mark standard plate to be under the camera mark standard plate according to the height value, namely, the coordinates of marking points in the ith radar mark plate are (x)_j,y_j0) into the coordinate system of the camera mark (x)_j,y_j+(h_li-h_c) 0) to obtain position prior information of each laser mark in a camera coordinate system;

step 6) inputting prior information based on a set sequence, and framing laser radar mark points (x) on laser radar visualization software according to the set sequence_l,y_l,z_l) The camera obtains the transformation relation between the camera mark coordinate system and the camera coordinate system through the prior size information of the camera mark]_m2cThereby marking the camera with coordinates (x) in the coordinate system_lm,y_lm,z_lm) Based on the formulaConverting the coordinate into a camera coordinate system to obtain the coordinate (x) of the laser radar mark point in the camera coordinate system_lc,y_lc,z_lc)；

Step 7) after the corresponding point information is obtained, the corresponding point information is obtained throughMinimize the reprojection error, i.e.:thereby obtaining the optimized rotation and translation relation between the laser radar and the camera coordinate system;

calibration board support on from top to bottom be equipped with through strong magnetism fixed laser mark standard board and camera mark standard board, wherein: the laser marking standard plate is arranged above the camera marking standard plate in a matrix manner;

the laser marking standard plate is provided with a guide strip, a marking point, a right side scale auxiliary line, a left side scale auxiliary line and a middle alignment auxiliary line;

the camera mark standard board is provided with an Aruco mark;

scales are arranged on two sides of the calibration plate bracket;

the laser radar visualization software is as follows: the control software RVIZ of the Robot Operating System (ROS).

2. The calibration method for the multi-line lidar and camera combined calibration according to claim 1, wherein the number of the lidar marking standard plates is preferably at least 20 laser points; the corresponding point information is preferably: the number of matched laser points is more than 20.

3. The method as claimed in claim 1, wherein the calibration module comprises: data acquisition unit, characteristic point extraction element and joint optimization unit, wherein: the data acquisition unit is connected with the camera, collects spatial perception information and outputs the spatial perception information to the feature extraction unit, the feature extraction unit screens out corresponding spatial laser radar feature information from the spatial perception information according to different radar point intensity information according to a laser radar feature extraction algorithm and corresponds to the visual feature point information one by one, and the combined optimization unit calculates external parameters among different sensors according to the spatial laser radar feature information and the corresponding visual feature point information.

Technical Field

The invention relates to a technology in the field of sensor combined calibration, in particular to a calibration method for combined calibration of a multi-line laser radar and a camera

Background

In the existing method for jointly calibrating the static laser radar and the camera, edge points are searched through the distance jump of the laser radar, the vertex of the calibration plate is fitted, the camera obtains corresponding vertex coordinates through identifying the mark of the camera and prior information, the problem that the edge point laser radar points have large errors, and the errors can be enlarged through line fitting exists; and a laser radar calibration mark is randomly placed around the camera mark, the position information of the prior characteristic point is obtained by measuring the relative position, and the radar characteristic point is found out by laser reflection intensity.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides a calibration method for the multi-line laser radar and camera combined calibration, and the universality and the usability of the calibration method are improved by standardizing radar marks and carrying out modular splitting and dynamic reconstruction on a calibration plate.

The invention is realized by the following technical scheme:

the invention relates to a calibration method for multi-line laser radar and camera combined calibration, which comprises the following steps:

step 1) moving a calibration plate bracket until the laser radar point is coincident with the plane of the Aruco pixel, adjusting the posture and leveling;

calibration board support on from top to bottom be equipped with through strong magnetism fixed laser mark standard board and camera mark standard board, wherein: the laser marking standard plate is arranged above the camera marking standard plate in a matrix manner.

The laser marking standard plate is provided with a guide strip, a marking point, a right side scale auxiliary line, a left side scale auxiliary line and a middle alignment auxiliary line.

The camera mark standard board is provided with an Aruco mark.

And scales are arranged on two sides of the calibration plate bracket.

The bottom of the calibration plate bracket is provided with a moving base with a roller, and the roller has leveling and locking functions.

The motion base is provided with a level gauge.

Step 2) adsorbing the camera marking standard plate on a calibration plate support, calibrating internal parameters of the camera, inputting a calibration result into a calibration module, adjusting the pose of the camera marking standard plate, adjusting the front-back distance of the calibration plate support, and locking the frame after a proper position is found;

step 3) opening laser radar visualization software, adjusting the upper and lower positions of the laser radar marking standard plate according to the reflection intensity, and placing the laser radar marking standard plate;

the laser radar visualization software is as follows: the control software RVIZ of the Robot Operating System (ROS).

The number of the laser radar marking standard plate is preferably at least 20 laser points.

Step 4) reading corresponding height value h of each mark standard plate_c,h_l1,h_l2,h_l3,...；

Step 5) converting points on the radar mark standard plate to be under the camera mark standard plate according to the height value, namely, the coordinates of marking points in the ith radar mark plate are (x)_j,y_j0) into the coordinate system of the camera mark (x)_j,y_j+(h_li-h_c) 0) to obtain position prior information of each laser mark in a camera coordinate system;

step 6) inputting prior information based on a set sequence, and framing laser radar mark points (x) on laser radar visualization software according to the set sequence_l,y_l,z_l) The camera obtains the transformation relation between the camera mark coordinate system and the camera coordinate system through the prior size information of the camera mark]_m2cThereby marking the camera with coordinates (x) in the coordinate system_lm,y_lm,z_lm) Based on the formulaConverting the coordinate into a camera coordinate system to obtain the coordinate (x) of the laser radar mark point in the camera coordinate system_lc,y_lc,z_lc)；

Step 7) after the corresponding point information is obtained, the corresponding point information is obtained throughMinimize the reprojection error, i.e.:thereby obtaining the optimized rotation and translation relation between the laser radar and the camera coordinate system.

The corresponding point information is preferably: the number of matched laser points is more than 20.

Technical effects

The invention integrally solves the technical problem that the matching relation between the laser point information and the pixel information is difficult to find in the existing calibration process; compared with the prior art, the calibration board can realize the association matching of the pixel plane and the laser point pair at one time, and simplify the matching process of the laser point information and the pixel point information.

Drawings

FIG. 1 is a schematic structural diagram of an improved calibration plate for multi-line lidar and camera joint calibration;

FIG. 2 is a calibration process for multi-line lidar and camera joint calibration;

FIG. 3 is a schematic illustration of a calibration plate;

FIG. 4 is a diagram illustrating an exemplary scenario;

FIG. 5 is a schematic flow chart of an embodiment;

in the figure: the laser leveling and locking device comprises a laser marking standard plate 1, a guide bar 2, marking points 3, a right side scale auxiliary line 4, a left side scale auxiliary line 5, a middle alignment auxiliary line 6, a camera marking standard plate 7, an ArUco marking point 8, a calibration plate support 9, a scale 10, a strong magnet 11, a motion base 12, a roller 13 with leveling and locking functions and a level meter 14.

Detailed Description

As shown in fig. 4, this embodiment is implemented by jointly calibrating a commercially available 16-line lidar and a camera, where the distance between the two sensors is relatively large, and the camera is located at a position about 60cm in front of the lidar and 50cm below the lidar.

As shown in fig. 1, the calibration board for the combined calibration of the multiline lidar and the camera according to the embodiment includes: calibration plate support 9 and from top to bottom set up in wherein through strong magnetism 11 fixed laser mark standard board 1 and camera mark standard board 7, wherein: the laser marking standard plate 1 is disposed above the camera marking standard plate 7 in a matrix manner.

The laser marking standard plate 1 is provided with a guide strip 2, a marking point 3, a right side scale auxiliary line 4, a left side scale auxiliary line 5 and a middle alignment auxiliary line 6.

The camera mark standard board 7 is provided with an Aruco mark 8.

And graduated scales 10 are arranged on two sides of the calibration plate bracket 9.

The bottom of the calibration plate bracket 9 is provided with a moving base 12 with a roller 13, and the roller 13 has leveling and locking functions.

The motion base 12 is provided with a level 14.

As shown in fig. 5, the present embodiment relates to a calibration method of the above apparatus, including the following steps:

the method comprises the following steps: firstly, completing the calibration of the internal parameters of the camera and inputting the calibration result into a calibration module;

step two: using an Aruco mark as a camera mark standard plate, placing the camera mark standard plate on a calibration plate support, selecting a corresponding standard plate model, opening a camera calibration module, moving the camera mark standard plate up and down, leveling and aligning an alignment center line and left and right scales, and moving the camera calibration plate support back and forth to enable a camera to stably identify the mark and enable a coordinate axis to point stably;

step three: after the camera is moved to a proper position, the roller of the camera calibration plate bracket is locked;

step four: opening control software RVIZ of a Robot Operating System (ROS), sequentially placing laser radar marking standard plates facing a visualization tool, and finishing leveling and aligning;

step five: reading the height value of each marking standard plate, obtaining calibration plate prior information based on the model and the height value of each plate, and inputting the calibration plate prior information into a calibration module according to a certain sequence;

step six: starting a calibration module, and performing frame selection on the marking points of the laser radar according to the sequence recorded in the step five;

step seven: operating calibration, projecting the calibration result to evaluate the calibration effect, and if the superposition of the laser radar point and the camera image is good in projection, considering that the calibration result is usable, otherwise, checking problems in the flow and re-calibrating;

the calibration module comprises: data acquisition unit, characteristic point extraction element and joint optimization unit, wherein: the data acquisition unit is connected with the camera, collects spatial perception information and outputs the spatial perception information to the feature extraction unit, the feature extraction unit screens out corresponding spatial laser radar feature information from the spatial perception information according to different radar point intensity information according to a laser radar feature extraction algorithm and corresponds to the visual feature point information one by one, and the combined optimization unit calculates external parameters among different sensors according to the spatial laser radar feature information and the corresponding visual feature point information.

Compared with the prior art, the device greatly optimizes the required operation time and applicability of the off-line calibration method, can smoothly adapt to the calibration of external parameters of various poses of the camera and the laser radar within a certain range, does not need to manufacture a calibration plate again for larger position change, forms a relatively standardized operation flow, reduces the operation difficulty, greatly reduces the dependence on operation experience, and greatly shortens the time for placing the calibration plate.

The foregoing embodiments may be modified in many different ways by those skilled in the art without departing from the spirit and scope of the invention, which is defined by the appended claims and all changes that come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.