Portable multi-sensor joint automatic calibration method

Abstract

The application discloses a portable multi-sensor combined automatic calibration method, belongs to the automatic driving technical field, and is suitable for multi-sensor combined automatic calibration in a non-fixed scene. The application combines a hardware platform design and a software calibration program, pushes out a portable calibration platform hardware device and multi-sensor combined automatic calibration, increases a cooperative mechanical arm hardware, and makes the multi-sensor combined calibration process automatic. After accurate pose control of the mechanical arm, the calibration process increases relative pose constraints, and the calculation amount is reduced through a RANSAC method, the optimization efficiency is improved, finally, the sensor external parameter calibration precision is improved through the portable calibration platform hardware setting, and the whole calibration process is simple, easy to use and efficient.

Description

Technical Field

[0001] The present invention belongs to the field of autonomous driving technology, and specifically relates to a portable multi-sensor joint automatic calibration method, which is suitable for multi-sensor joint automatic calibration in non-fixed scenarios. Background Art

[0002] In the existing field of multi-sensor joint calibration for autonomous driving, due to the fact that different vehicles have assembly errors in the installation positions of sensors before leaving the factory, it is impossible to achieve complete uniformity. In addition, there will be certain errors in the production and assembly process of sensor components. Therefore, multiple sensors must undergo unified data collection and calibration in the calibration workshop before leaving the factory. Common joint calibration methods are generally the following two:

[0003] The first is the markerless calibration method. This method does not require a specific calibration plate or artificially constructed scenes. It generally performs automated internal and external parameter calibration in natural outdoor scenes, but the calibration process requires the vehicle to remain in motion. The advantage of this method is its wide range of scene freedom and flexible calibration timing. However, its disadvantage is that natural scenes sometimes lack obvious characteristic markers, making calibration accuracy difficult to guarantee.

[0004] The second method is the calibration method using markers in specific locations. In specific scenarios, such as a calibration workshop, the vehicle is parked in a specific central location or on a rotating platform. Special calibration plates are affixed to the surrounding area in a certain pattern, or multiple robotic arms are suspended from the beam arm. A special calibration plate is fixed to the end of each robotic arm. The program pre-sets the robotic arm's trajectory, then starts the program to collect camera and lidar data on the vehicle. Finally, the calibration algorithm is run to obtain the internal and external parameters of the multiple sensors. The advantage of this method is that because the calibration workshop is a relatively fixed environment, higher-precision and more consistent sensor calibration can be achieved. For the mass production of smart driving vehicles, calibration workshops are more convenient for forming standardized operating procedures. However, this method has strict requirements on the calibration site, and once the user replaces or reinstalls a single sensor during use, the entire vehicle must be returned to the factory for recalibration. This is a long cycle and has a high time cost for customers. Especially for remote areas with weak infrastructure, the calibration and maintenance of multiple sensors on the entire vehicle is very difficult.

[0005] The innovation of this invention lies in the fact that the calibration platform hardware device is lightweight and portable, easy to disassemble and install, and can automatically carry out the multi-sensor calibration process in non-fixed scenarios with guaranteed calibration accuracy, greatly shortening the calibration cycle and reducing customer maintenance costs. Summary of the Invention

[0006] (1) Purpose of the invention

[0007] The application aims to provide a portable multi-sensor joint automatic calibration method, which combines hardware platform design and software calibration program, and pushes out a portable calibration platform hardware device and multi-sensor joint automatic calibration.

[0008] (II) Technical scheme

[0009] A portable multi-sensor joint automatic calibration method is realized by combining the portable calibration platform hardware device and the multi-sensor joint automatic calibration program through the combination of software and hardware.

[0010] The portable calibration platform hardware device comprises a tire fixator, a scissor type endless telescopic arm and a six-degree-of-freedom mechanical arm.

[0011] The tire fixator adopts a double-screw clamp block structure and can be fixed with high precision according to different sizes of tires.

[0012] The scissor type endless telescopic arm comprises an angle plate, a scissor type telescopic structure, a mechanical arm fixing plate and a laser ranging and receiving module.

[0013] The second laser ranging and receiving module is installed on the mechanical arm fixing plate and is used to return the relative position data of the one-axis center of the mechanical arm to the wheel to the control center for later calibration.

[0014] The six-degree-of-freedom mechanical arm is used to meet the calibration requirements according to the specific size of the vehicle to be calibrated.

[0015] During calibration, one or two sets of tire fixators are installed on the same side of the front and rear tires of the vehicle to be calibrated.

[0016] Adjust the front and rear wheels of the vehicle to the same level according to the prompts of the laser ranging and receiving module;

[0017] Select one or more sets of scissor-type telescopic arms and motion execution terminals according to the calibration requirements, and complete data collection in the area covered by the corresponding calibration board by adjusting the installation positions of the scissor-type telescopic arms and motion execution terminals;

[0018] Start the multi-sensor system. The six-degree-of-freedom robotic arm will move along the preset trajectory in conjunction with the calibration plate. Start the multi-sensor joint automatic calibration program to automatically collect multi-frame time-synchronized robotic arm pose and multi-sensor data. Start automatic calibration of the multi-sensor internal and external parameters. Repeat the six-degree-of-freedom robotic arm operation until the accuracy reaches the required level.

[0019] The multi-sensor automatic calibration program completes the following actions:

[0020] Step 1, preparation before calibration;

[0021] Adjust the laser rangefinder so that the vehicle body, the connecting rod of the scissor-type telescopic structure, and the six-degree-of-freedom robotic arm are in a horizontal state:

[0022] Step 2: Calibrate the camera’s intrinsic parameters and the camera and lidar’s extrinsic parameters.

[0023] The relative pose constraints of the six-degree-of-freedom manipulator are introduced to improve the calibration accuracy, and a random sampling method is used to reduce the number of relative pose constraints and improve the calculation efficiency:

[0024] Step 3: Calibrate the external parameters of the laser radar relative to the vehicle body;

[0025] By presetting the robotic arm posture, the calibration plate and the robotic arm base are successively in a horizontal and vertical state, thereby obtaining the transformation matrix from the lidar coordinate system to the vehicle coordinate system, that is, the external parameters of the lidar relative to the vehicle body.

[0026] (3) Effective income

[0027] 1. The present invention is lightweight and portable, modular in structure, and easy to disassemble and install;

[0028] 2. The present invention has wide adaptability, strong scalability of the hardware platform, and modular scalability, making it adaptable to various types and sizes of autonomous vehicles, such as trucks and cars.

[0029] 3. The present invention is mobile and flexible, not limited to factory buildings, and can be adapted to remote areas with weak infrastructure;

[0030] 4. After the present invention introduces the robotic arm, its high-precision pose information can provide constraints on relative pose transformation, thereby improving the accuracy of the extrinsic parameters between the camera and the lidar;

[0031] 5. After adding the pose information of the robotic arm, the present invention can obtain the precise transformation of the laser radar relative to the vehicle coordinate system based on the coordinate relationship designed by the hardware platform, that is, the transformation from the laser radar coordinate system to the vehicle coordinate system;

[0032] 6. The overall cost of the present invention is lower than that of factory repair, and the cycle is short, the operation is simple, and it is easy to implement. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 Working diagram of the portable calibration platform of the present invention;

[0034] Figure 2 Schematic diagram of the working process of the present invention;

[0035] Figure 3 A schematic diagram of the tire fixer of the present invention;

[0036] Among them: 11- double screw clamp structure, 12- circular ear structure, 13- first laser ranging and receiving module

[0037] Figure 4 Schematic diagram of the scissor-type infinitely telescopic arm of the present invention;

[0038] Among them: 31-angle plate, 32-scissor-type telescopic structure, 33-mechanical arm fixing plate, 34-second laser ranging and receiving module

[0039] Figure 5 Schematic diagram of the six-degree-of-freedom robotic arm of the present invention;

[0040] Figure 6 Schematic diagram of the combined calibration plate of the present invention. DETAILED DESCRIPTION

[0041] The present invention will be explained and illustrated in detail below with reference to the accompanying drawings and embodiments.

[0042] The multi-sensor joint automatic calibration referred to in this invention mainly refers to the most common calibration between cameras and lidars. The calibration objects mainly include the intrinsic parameters of the camera, the extrinsic parameters of the camera and radar, and the extrinsic parameters from the sensor to the vehicle body.

[0043] The portable multi-sensor joint automatic calibration method of the present invention is implemented through a combination of hardware and software: the construction of a portable calibration platform hardware device and the implementation of a multi-sensor joint automatic calibration program. This method can solve the problem of multi-sensor joint automatic calibration for autonomous vehicles in non-stationary scenarios. Compared with existing calibration methods that require fixed locations, this method offers the advantages of greater flexibility and convenience.

[0044] (1) Hardware setup of portable calibration platform

[0045] The portable calibration platform hardware device of the present invention is as follows: Figure 1 As shown, it includes two tire fixers 1, a scissor-type infinitely telescopic arm 3, and a set of six-degree-of-freedom robotic arms 4. Each component uses a unified mechanical interface, and different components can be selected according to actual needs, not limited to the number of the above components. The tire fixer 1 adopts a double-screw clamping block structure 11, as shown in FIG. Figure 3 As shown, it can be installed and fixed with high precision according to different tire sizes. It has a built-in first laser ranging and receiving module 13 to ensure that the front and rear wheels are facing the same direction as the vehicle body. The principle is that the laser module on one tire holder illuminates the laser receiver on the other tire holder to adjust the steering wheel. When the laser receiving module reads correctly, it indicates that the front and rear wheels are facing the same direction as the vehicle body, which is conducive to the calibration of external parameters of multiple sensors to the vehicle body coordinate system. A circular hanging ear 12 is designed on the outside of the cover of the first laser ranging and receiving module, which enables the tire holder 1 to be installed in a horizontal or vertical posture. The scissor-type infinitely telescopic arm 3 is composed of four parts: an angle plate 31, a scissor-type telescopic structure 32, a mechanical arm fixing plate 33, and a laser ranging and receiving module 34. Figure 4 As shown. The angle plate 31 can be fixedly connected to the circular ear structure 12 in the tire holder 1. The angle plate 31 is designed with limiting holes at several angles such as 90°, 45°, and 0°. It can fix the scissor-type telescopic arm at different angles according to actual calibration requirements, so as to determine the relative position of the robot arm and the tire holder.

[0046] The scissor-type telescopic structure 32 utilizes a hollow cable routing design made of 7075 aviation aluminum, with wire troughs located on the inside of each connecting rod, ensuring the safety of the wires without compromising telescoping. A second laser ranging and receiving module 34 is installed on the robotic arm mounting plate 33. This module transmits the relative position data from the center of the robotic arm's axis to the wheel back to the control center for later calibration. This module measures the relative distance between the robotic arm mounting plate's base coordinates and the vehicle body, facilitating calibration of the sensor and vehicle's external parameters. A mechanical mounting interface for the six-degree-of-freedom robotic arm 4 is reserved on the robotic arm mounting plate 33, enabling modular installation of the robotic arm.

[0047] The six-degree-of-freedom manipulator 4 selects a suitable manipulator model according to the specific size of the vehicle 2 to be calibrated to meet the calibration requirements. A joint calibration plate 5 is provided at the end of the six-degree-of-freedom manipulator 4, such as Figure 5 、 6 As shown, the joint calibration board is used for joint calibration of sensors such as cameras and lidars. The four corners of the calibration board are hollow circles, and the middle area is a common checkerboard image. The calibration board is hard-connected to the end of the six-degree-of-freedom robotic arm 4.

[0048] The calibration process of the portable calibration platform hardware device is as follows: Figure 2 shown.

[0049] First, install a set or two sets of tire fixers on the same side of the front and rear tires of the vehicle to be calibrated, and connect the wire harness to the control center device and power on.

[0050] According to the prompt of the laser ranging and receiving module, adjust the front and rear wheels of the vehicle to the same horizontal plane.

[0051] According to the calibration requirements, install one or more sets of scissor telescopic arms and action execution ends, and complete the data acquisition of the corresponding calibration plate coverage area by adjusting the installation positions of the scissor telescopic arms 3 and the action execution ends, such as left front, right front, left rear and right rear.

[0052] Start the multi-sensor six-degree-of-freedom robot arm 4 to move the joint calibration plate 5 according to the preset trajectory, start the multi-sensor joint automatic calibration program to automatically collect multiple frames of time-synchronized robot arm poses and multi-sensor data, and start automatic calibration of the multi-sensor internal and external parameters. If the calibration accuracy is not reached, repeat the operation of the six-degree-of-freedom robot arm 4. If the calibration reaches the accuracy, the calibration process is completed, and the six-degree-of-freedom robot arm stops at the initial position.

[0053] (II) Multi-sensor joint automatic calibration program

[0054] For the joint calibration problem of typical camera and laser radar sensors of an autonomous vehicle, the main solutions are camera internal parameters, camera coordinate system to laser radar coordinate system external parameters, and laser radar coordinate system to vehicle body coordinate system external parameters. The automatic calibration method mainly corresponds Figure 2 Collect multiple frames of robot arm poses and multi-sensor data, and start automatic calibration. Note that the robot arm end is hard connected with the calibration plate, so the pose of the calibration plate can be obtained through the pose of the robot arm. Therefore, in the description of the calibration method, the pose data of the calibration plate or the robot arm can be regarded as the same kind of data unless specifically stated.

[0055] The automatic calibration method adopted by the present application mainly includes three parts.

[0056] The first part is the preparation before calibration. Adjust the laser range finder so that the vehicle body, connecting rod and robot arm module are in a horizontal state as much as possible:

[0057] In this method, first, drive the vehicle to a flat ground as much as possible, not limited to indoor or outdoor, adjust the rear wheel laser ranging module so that it can emit a signal to the front wheel laser receiving module, set the signal feedback receiving success, which indicates that the front and rear wheels are adjusted to be consistent with the vehicle body; At the same time, adjust the pose of the robot arm fixing plate so that its laser receiving module can receive the front wheel laser emission signal, which indicates that the robot arm fixing plate is also horizontal with the vehicle body, and the next step of calibration operation can be continued.

[0058] The second part involves calibrating the camera's intrinsic parameters and the camera and lidar's extrinsic parameters. The relative pose constraints of the robotic arm are introduced to improve calibration accuracy, and the RANSAC (Random Sample Consensus) method is used to reduce the number of relative pose constraints and improve computational efficiency.

[0059] When calculating the relative transformation projection error, considering that the introduction of the relative posture transformation constraint of the manipulator can improve the calibration accuracy, but the amount of calculation also increases sharply, the random sample consensus (RANSAC) method is adopted, that is, some relative posture constraints are randomly extracted in an iterative manner to optimize the solution variables, thereby achieving the purpose of improving computational efficiency.

[0060] The third part is the calibration of the extrinsic parameters of the LiDAR relative to the vehicle body. By presetting the robotic arm posture, the calibration plate and the robotic arm base are placed in a horizontal and vertical state, thereby obtaining the transformation matrix from the LiDAR coordinate system to the vehicle coordinate system, that is, the extrinsic parameters of the LiDAR relative to the vehicle body:

[0061] The main steps are as follows:

[0062] 1. According to the preset procedure of the LiDAR external parameter calibration, the robotic arm automatically adjusts its posture so that the calibration plate and the robotic arm base are horizontal, and the LiDAR point cloud data is collected at this time;

[0063] 2. According to the method for calculating the 3D coordinates of the center of the hollow circle in the lidar point cloud in the second part, the 3D coordinates of the center of the circle are calculated, recorded as O = [x0, y0, z0], and the point cloud data of the calibration plate area in the lidar point cloud are extracted, and the unit normal vector of the plane point cloud is estimated, recorded as

[0064] 3. Adjust the robot arm posture so that the calibration plate and the robot arm base are perpendicular, and then collect the lidar point cloud data again;

[0065] 4. Extract the point cloud data of the calibration plate area in the lidar point cloud and estimate the unit normal vector of the plane point cloud, which is recorded as

[0066] 5. At this time, take O=[x0,y0,z0] as the coordinate origin, as well as and The cross product of are three coordinate axes, forming the calibration plate coordinate system O b , its pose relative to the lidar coordinate system is expressed as where R LO and t LOrepresents the rotation and translation of the calibration plate coordinate system relative to the lidar coordinate system, P LO It is a 4×4 matrix, which represents the transformation from the calibration plate coordinate system to the lidar coordinate system;

[0067] 6. Based on the mechanical dimensions of the calibration plate and the geometric relationship between the robotic arm, the relative transformation of the calibration plate center coordinates to the robotic arm base can be calculated. Based on the structure of the portable calibration platform hardware device in this method, the transformation of the calibration plate coordinate system relative to the vehicle coordinate system can be obtained, which is recorded as where R VO and t VO represents the rotation and translation of the calibration plate coordinate system relative to the lidar coordinate system, P VO Indicates the transformation from the calibration plate coordinate system to the vehicle coordinate system;

[0068] 7. Let the transformation matrix from the laser radar coordinate system to the vehicle coordinate system be P VL , from the relative transformation relationship of the coordinate system, we can know that in Represents the transformation matrix P LO The inverse of . So far, we get the external parameter calibration P of the variable laser radar relative to the vehicle body. VL .

[0069] In summary, through the three-part calibration process, the calibration of camera intrinsic parameters, extrinsic parameters from the camera coordinate system to the lidar coordinate system, and extrinsic parameters from the lidar coordinate system to the vehicle coordinate system is completed.

[0070] The advantage of the method of the present invention is that after the introduction of precise posture control of the robotic arm, relative posture constraints are added to the calibration process, and the RANSAC method is used to reduce the amount of calculation and improve the optimization efficiency. Finally, the hardware setting of the portable calibration platform is used to improve the calibration accuracy of the sensor external parameters, making the entire calibration process simple, easy to use and efficient.

[0071] The above content is a further detailed description of the present invention in conjunction with specific embodiments, and the specific implementation of the present invention should not be considered to be limited to these descriptions. For those skilled in the art of the present invention, without departing from the concept of the present invention, several simple deductions or substitutions can be made, which should be considered to fall within the scope of protection of the present invention.

Claims

1. A portable multi-sensor joint automatic calibration method, characterized in that: This is achieved through the combination of hardware and software, namely the construction of a portable calibration platform hardware device and a multi-sensor joint automatic calibration program; The portable calibration platform hardware device includes a tire holder, a scissor-type infinitely telescopic arm, and a six-degree-of-freedom robotic arm. Each component uses a unified mechanical interface, and different components can be selected according to actual needs. The number of these components can be selected according to specific task requirements. The tire holder uses a dual-screw clamp structure, enabling high-precision installation and fixation of tires of different sizes. A built-in laser ranging and receiving module ensures that the front and rear wheels are aligned with the vehicle's orientation. Round lugs are designed on the outside of the cover of the module, enabling the tire holder to be installed in either a horizontal or vertical position with the angle plate of the scissor-type infinitely telescopic arm. The scissor-type infinitely telescopic arm consists of an angle plate, a scissor-type telescopic structure, a robotic arm fixing plate, and a laser ranging and receiving module. The angle plate is fixedly connected to the circular hanging ear structure. The angle plate is designed with limit holes at 90°, 45°, and 0° to fix the scissor-type infinitely telescopic arm at different angles according to actual calibration requirements, so as to determine the relative position of the robotic arm and the tire holder. A second laser ranging and receiving module is installed on the robot arm fixing plate to transmit the relative position data from the center of the robot arm's axis to the wheel back to the control center for later calibration. The robot arm fixing plate also reserves a mechanical installation interface for the six-degree-of-freedom robot arm, enabling modular installation of the robot arm. The 6DOF manipulator selects the appropriate manipulator model based on the specific size of the vehicle to be calibrated to meet the calibration requirements; a joint calibration plate is set at the end of the 6DOF manipulator for the joint calibration of the camera and lidar sensor; The calibration process of the portable calibration platform hardware device is as follows: Step 1: Install one or two sets of tire holders on the front and rear tires on the same side of the vehicle to be calibrated, connect the wiring harness to the control center equipment and power on; Step 2: Adjust the front and rear wheels of the vehicle to the same level according to the prompts of the laser ranging and receiving module; Step 3: Select and install one or more sets of scissor-type infinitely telescopic arms and motion execution terminals according to the calibration requirements, and complete data collection in the area covered by the corresponding calibration board by adjusting the installation positions of the scissor-type infinitely telescopic arms and motion execution terminals; Step 4: Start the multi-sensor system. The six-degree-of-freedom robotic arm will move along the preset trajectory in conjunction with the calibration plate. Start the multi-sensor joint automatic calibration program to automatically collect multi-frame time-synchronized robotic arm pose and multi-sensor data. Start automatic calibration of the multi-sensor internal and external parameters. Repeat the six-degree-of-freedom robotic arm operation until the accuracy reaches the required level. The multi-sensor combined automatic calibration procedure completes the following actions: Step 1, preparation before calibration; Adjust the laser rangefinder so that the vehicle body, the connecting rod of the scissor-type telescopic structure, and the six-degree-of-freedom robotic arm are in a horizontal state: Step 2: Calibrate the camera’s intrinsic parameters and the camera and lidar’s extrinsic parameters. The relative pose constraints of the six-degree-of-freedom manipulator are introduced to improve the calibration accuracy, and a random sampling method is used to reduce the number of relative pose constraints and improve the calculation efficiency: Step 3: Calibrate the external parameters of the laser radar relative to the vehicle body; By presetting the robotic arm posture, the calibration plate and the robotic arm base are successively in a horizontal and vertical state, thereby obtaining the transformation matrix from the lidar coordinate system to the vehicle coordinate system, that is, the external parameters of the lidar relative to the vehicle body.

2. A portable multi-sensor joint automatic calibration method according to claim 1, characterized in that: Step 3 of the multi-sensor joint automatic calibration procedure specifically includes the following steps: 3.1 According to the preset procedure of the LiDAR external parameter calibration, the robotic arm automatically adjusts its posture so that the calibration plate and the robotic arm base are in a horizontal state, and the LiDAR point cloud data is collected at this time; 3.2 According to the method of the three-dimensional coordinates of the center of the hollow circle in the lidar point cloud, the three-dimensional coordinates of the circle center are calculated, recorded as O = [x0, y0, z0], and the point cloud data of the calibration plate area in the lidar point cloud are extracted, and the unit normal vector of the plane point cloud is estimated, recorded as 3.3 Adjust the robot arm posture so that the calibration plate and the robot arm base are perpendicular to each other, and collect the lidar point cloud data again; 3.4 Extract the point cloud data of the calibration plate area in the lidar point cloud and estimate the unit normal vector of the plane point cloud, which is recorded as At this time, take O=[x0,y0,z0] as the coordinate origin, as well as and The cross product of are three coordinate axes, forming the calibration plate coordinate system O b , its pose relative to the lidar coordinate system is expressed as where R LO and t LO represents the rotation and translation of the calibration plate coordinate system relative to the lidar coordinate system, P LO It is a 4×4 matrix, which represents the transformation from the calibration plate coordinate system to the lidar coordinate system; 3.5 According to the mechanical dimensions of the calibration plate and the geometric relationship between the manipulator, the relative transformation of the calibration plate center coordinates to the manipulator base is calculated. According to the structure of the portable calibration platform hardware device, the transformation of the calibration plate coordinate system relative to the vehicle coordinate system is obtained, which is recorded as where R VO and t VO represents the rotation and translation of the calibration plate coordinate system relative to the lidar coordinate system, P VO Indicates the transformation from the calibration plate coordinate system to the vehicle coordinate system; 3.6 Let the transformation matrix from the laser radar coordinate system to the vehicle coordinate system be P VL , from the relative transformation relationship of the coordinate system, we can know that in Represents the transformation matrix P LO The inverse of the variable, so far, the external parameter calibration P of the laser radar relative to the vehicle body is obtained. VL .

3. A portable multi-sensor joint automatic calibration method according to claim 1 is characterized in that the scissor-type telescopic structure adopts a hollow wiring design of 7075 aviation aluminum, and the wire groove is designed on the inner side of each connecting rod, ensuring the safety of the wire without affecting the telescopic structure.

4. A portable multi-sensor joint automatic calibration method according to claim 1, characterized in that: The four corners of the calibration plate are hollow circles, the middle area is a common checkerboard image, and the calibration plate is hard-connected to the end of the six-degree-of-freedom robotic arm.

5. The portable multi-sensor joint automatic calibration method according to claim 1, characterized in that: Step 1 of the multi-sensor joint automatic calibration procedure is as follows: drive the vehicle onto a flat surface, whether indoors or outdoors, adjust the rear wheel laser ranging module so that it can transmit signals to the front wheel laser receiving module, and set the signal feedback reception to be successful, indicating that the front and rear wheels are adjusted to be consistent with the vehicle body; at the same time, adjust the posture of the robot arm fixing plate so that the laser receiving module can receive the front wheel laser emission signal, indicating that the robot arm fixing plate is also level with the vehicle body and can proceed to the next calibration operation.

6. A portable multi-sensor joint automatic calibration method according to claim 1, characterized in that: In step 2 of the multi-sensor joint automatic calibration procedure, when calculating the relative transformation projection error, a random sampling consensus method is used to iteratively randomly extract some relative pose constraints to optimize the solution variables, thereby achieving the purpose of improving computational efficiency.

Images