Linear scanning pose calibration system and method based on ChArUco calibration plate
By using ChArUco calibration plate and RefineNet neural network combined with PnP algorithm on an online laser scanner, the all-round high-precision posture calibration of the linear laser scanner is achieved, solving the problems of low calibration accuracy and complex operation in the existing technology. It is suitable for Multiple scenarios and reduce system costs.
Patent Information
- Application Number
- CN202410940778.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-12
AI Technical Summary
The prior art is difficult to realize simple and effective posture calibration of line laser scanners, with low calibration accuracy, complex operation, and not suitable for all-round multi-scene.
Using a line scanning posture calibration system based on ChArUco calibration plate, a line laser scanner is installed on a rotatable tripod gimbal, and a ChArUco calibration plate is fixedly placed every 120 degrees. Combined with the RefineNet neural network and PnP algorithm, the line laser scanning is solved. The pose relationship between the coordinate system of the instrument and the world coordinate system is carried out, and multiple sets of data iterative optimization is carried out to obtain the most accurate pose.
It realizes all-round high-precision line laser scanner position calibration, reduces system production costs, simplifies the operation process, is suitable for multiple scenarios, and reduces the impact of external environment interference on calibration accuracy.
Smart Images

Figure CN118941642A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of computer vision, and in particular relates to a line scan pose calibration system and method based on a ChArUco calibration board. Background Art
[0002] As deep learning and computer vision technology continue to develop in the direction of digitalization, intelligence, and convenience, non-contact three-dimensional weighing has become the key to industrial manufacturing and is widely used in the reconstruction of digital cities, damage detection of large-scale equipment, protection and restoration of monuments, part defect detection and shape reconstruction, and other fields. The most critical step in the process of achieving three-dimensional measurement is visual calibration technology. The accuracy of camera (scanner) calibration and the determination of intrinsic matrix, extrinsic matrix, distortion coefficient, and light plane matrix directly affect the quality of imaging. In recent years, with the continuous development of technology, advanced line laser scanners have gradually replaced the status of traditional industrial cameras. Line laser scanners can directly obtain three-dimensional shape information on the surface of objects, achieve high-precision measurement, are not limited by lighting conditions, have a large measurement range, use non-contact measurement methods, and also have the ability to automatically scan and process data. In summary, line laser scanners have more advantages than traditional industrial cameras.
[0003] At present, the calibration method for line laser scanners mainly adopts Zhang's "chessboard" calibration method. The calibration plate adopts a chessboard calibration plate or a Halcon (7*7) circular calibration plate, and multiple calibration plates with different postures are placed. Multiple groups of calibration plate pictures are taken with the laser turned on and off respectively. The corner points are extracted from the multiple pictures taken with the laser turned off, and the internal and external parameters of the camera are calculated. The laser line pixel coordinates are extracted from the multiple pictures taken with the laser turned on. The camera internal and external parameters calculated in the above steps are used to solve the light plane matrix. However, Zhang's "chessboard" calibration method has high requirements on the quality of the calibration plate and is limited to the calibration of a single plane. It is not suitable for calibration plates of very large sizes. It may be necessary to combine other calibration methods or take special measures to overcome these disadvantages.
[0004] The invention patent "Joint online calibration method of inertial and laser scanners" with application publication number CN114162169A makes and obtains the standard profile of the standard track model, and then places the track inspection beam on a two-degree-of-freedom swing table and adjusts it to a horizontal state. Perform initial alignment on the inertial navigation system, and synchronously collect laser profile information and inertial navigation attitude information. Adjust the profile output by each laser scanner to nearly coincide with the standard profile, and complete the rough calibration of the inertial navigation system and laser scanner. Finally, perform fine calibration. Although the detection accuracy is improved compared to the traditional posture calibration method, the operation is very complicated and difficult, and the applicability to the scene is low. It is suitable for scanner calibration in specific scenes.
[0005] The invention patent "A robot hand-eye calibration method based on a line laser scanner" with application publication number CN112659112A uses the center of a standard ball on one side of the robot as a calibration point. While controlling the robot's movements, the line laser scanner obtains the position coordinates of the calibration point in its coordinate system in different postures. Then, the nonlinear optimization method is used to solve the objective function to obtain the hand-eye calibration matrix of the position relationship between the line laser scanner and the robot's end effector. This calibration method effectively avoids the problem that the laser plane must pass directly through the calibration point when calibrating the line laser scanner, but the calibration method of this scanner is limited to a single plane and is not suitable for calibration of the entire scene. In addition, the noise in the measurement data is eliminated through nonlinear optimization (least squares method), which cannot effectively solve the experimental data error.
[0006] Therefore, how to simply and effectively calibrate the position and posture of the line laser scanner with high calibration accuracy, low system manufacturing cost, low operation difficulty, and applicability to all-round and multi-scene scenarios is a key issue that needs to be solved urgently at this stage. Summary of the invention
[0007] The purpose of the present invention is to propose a line scanning pose calibration system and method based on a ChArUco calibration plate. The line laser scanner is installed on a rotatable tripod platform. A ChArUco calibration plate is fixedly placed every 120 degrees on the platform plane. The fixed angle of each rotation of the platform shaft is set to 6 degrees. The shaft rotates 360 degrees in total, 60 times, and each rotation is a group. Multiple groups of calibration plate images with different poses are taken. After completing a complete rotation, one group is selected to determine multiple groups of key corner points at different positions on the calibration plate, and the key corner points of the ChArUco calibration plate are located and the corner point coordinates are revised in combination with the RefineNet neural network, and the pose relationship between the coordinate system where the line laser scanner is located and the world coordinate system is solved in combination with the PnP (Perspective-n-Point) algorithm and the relevant parameters of the laser scanner. The pose relationship of the remaining groups is also settled according to the above method, and finally the pose relationship solved by all the above groups is optimized, so that the most accurate pose can be obtained, and the overall pose calibration process of the line laser scanner is completed.
[0008] To achieve the above-mentioned object, the present invention provides a line scanning posture calibration system based on a ChArUco calibration plate, comprising: a line laser scanner, a connecting device, a plurality of ChArUco calibration plates, a telescopic device, and a rotatable supporting device;
[0009] The telescopic device is fixed on the rotatable supporting device, and the line laser scanner is fixed on the other end of the telescopic device;
[0010] The ChArUco calibration plates are centered on the rotatable support device and are evenly distributed at preset angles between each other;
[0011] The line laser scanner collects image data of the ChArUco calibration plate through the rotation of the rotatable support device, and uploads the image data to the host computer through the connecting device for position and posture marking.
[0012] Optionally, the ChArUco calibration plate includes: a first ChArUco calibration plate, a second ChArUco calibration plate, and a third ChArUco calibration plate;
[0013] The first ChArUco calibration plate is fixedly installed corresponding to 120 degrees of the rotatable support device; the second ChArUco calibration plate is fixedly installed corresponding to 240 degrees of the rotatable support device; and the third ChArUco calibration plate is fixedly installed corresponding to 360 degrees of the rotatable support device.
[0014] Optionally, the ChArUco calibration plate is formed by fusion of a checkerboard calibration plate and an ArUco calibration plate.
[0015] The present invention also proposes a line scan pose calibration method based on a ChArUco calibration board, the method comprising:
[0016] Obtain the intrinsic reference data of the line laser scanner;
[0017] Collect image data of the ChArUco calibration plate;
[0018] Using the RefineNet network, key corner points of the image data are located;
[0019] Based on the positioned image data, combined with the PnP algorithm and the internal reference data of the line laser scanner, several sets of pose values between the coordinate system where the line laser scanner is located and the world coordinate system are obtained, and several sets of pose values are used for iterative optimization to obtain the final pose value.
[0020] Optionally, checkerboard and ArUco calibration plate corner points are selected as key corner points of the image data.
[0021] Optionally, obtaining the internal reference data of the line laser scanner includes:
[0022] Converting the world coordinate system of the line laser scanner into a camera coordinate system;
[0023] Convert the camera coordinate system into an image coordinate system;
[0024] Converting the image coordinate system into a pixel coordinate system;
[0025] The intrinsic parameter data is calculated by combining the camera coordinate system, the image coordinate system and the pixel coordinate system.
[0026] Optionally, using a RefineNet network to locate key corner points of the image data includes:
[0027] Annotate the key corner points of the collected ChArUco calibration plate image dataset;
[0028] The RefineNet network is trained based on an image dataset with key corner points annotated;
[0029] The trained RefineNet network is used to locate key corner points of the image data and revise the coordinates of the corner points.
[0030] Optionally, in the RefineNet network, a 4-level cascade structure is formed by 4 RefineNet units, and high-resolution prediction results are generated by using multi-level feature extraction and long-distance residual connections.
[0031] Optionally, the group of pose values is:
[0032]
[0033] Among them, R represents the rotation matrix, T represents the translation matrix, and f ab Represents the calibration result parameters, a=1, 2, 3, b=1, 2, 3, 4.
[0034] The present invention has the following beneficial effects:
[0035] The present invention can realize the omni-directional collection of multiple groups of ChArUco calibration plate image data information at different angles, and is less affected by the external environment. The key corner points of the ChArUco calibration plate are located and the corner point coordinates are revised through the RefineNet neural network, and the pose relationship between the coordinate system where the line laser scanner is located and the world coordinate system is solved in combination with the PnP algorithm and the relevant parameters of the line laser scanner. Through the continuous iterative optimization of multiple groups of data and , the most accurate pose can be finally obtained, completing the overall pose calibration process of the line laser scanner. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0037] Figure 1A schematic diagram of a posture calibration structure of a line scan posture calibration system and method based on a ChArUco calibration board according to an embodiment of the present invention;
[0038] Figure 2 Schematic diagram of the overall workflow of a line scan pose calibration method based on a ChArUco calibration board according to an embodiment of the present invention;
[0039] Figure 3 A schematic diagram of the RefineNet network structure of an embodiment of the present invention;
[0040] Figure 4 Schematic diagram of a ChArUco calibration board according to an embodiment of the present invention;
[0041] Figure 5 A schematic diagram of key point selection of the ChArUco calibration plate according to an embodiment of the present invention;
[0042] Figure 6 A schematic diagram of the internal parameter calibration of a line laser scanner according to an embodiment of the present invention;
[0043] Among them, 1—line laser scanner, 2—interface between line laser scanner and host computer, 3—ChArUco calibration board No. 1, 4—ChArUco calibration board No. 2, 5—ChArUco calibration board No. 3, 6—retractable line laser scanner support column, 7—fixed support base of line laser scanner with rotating scale. DETAILED DESCRIPTION
[0044] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0045] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0046] This embodiment proposes a line scanning pose calibration system based on a ChArUco calibration plate, which includes: a line laser scanner, a connecting device, a plurality of ChArUco calibration plates, a telescopic device, and a rotatable supporting device;
[0047] The telescopic device is fixed on the rotatable supporting device, and the line laser scanner is fixed on the other end of the telescopic device;
[0048] The ChArUco calibration plates are centered on the rotatable support device and are evenly distributed at preset angles between each other;
[0049] The line laser scanner collects the image data of the ChArUco calibration plate through the rotation of the rotatable support device, and uploads the image data to the host computer through the connection device for position and pose marking.
[0050] Further, the ChArUco calibration plate includes: a first ChArUco calibration plate, a second ChArUco calibration plate, and a third ChArUco calibration plate;
[0051] The first ChArUco calibration plate is fixedly installed corresponding to 120 degrees of the rotatable support device; the second ChArUco calibration plate is fixedly installed corresponding to 240 degrees of the rotatable support device; and the third ChArUco calibration plate is fixedly installed corresponding to 360 degrees of the rotatable support device.
[0052] Furthermore, the ChArUco calibration plate is composed of a fusion of a checkerboard calibration plate and an ArUco calibration plate.
[0053] Specifically, this embodiment provides Figures 1 to 5 As shown, a line scanning posture calibration system and method based on a ChArUco calibration plate include a line laser scanner 1, an interface 2 between the line laser scanner and a host computer, a No. 1 ChArUco calibration plate 3, a No. 2 ChArUco calibration plate 4, a No. 3 ChArUco calibration plate 5, a retractable line laser scanner support column 6, and a line laser scanner fixed support base 7 with a rotating scale.
[0054] The line laser scanner 1 is fixed on a retractable line laser scanner support column 6, and the other end is connected to the host computer through the line laser scanner and host computer interface 2.
[0055] The retractable line laser scanner support column 6 is fixed on a line laser scanner fixed support base 7 with a rotating scale, ensuring that clear images of multiple groups of ChArUco calibration plates can be obtained.
[0056] The No. 1 ChArUco calibration plate 3 is fixedly installed at 120 degrees corresponding to the fixed support base of the line laser scanner with a rotating scale.
[0057] The No. 2 ChArUco calibration plate 4 is fixedly installed at 240 degrees corresponding to the fixed support base of the line laser scanner with a rotating scale.
[0058] The No. 3 ChArUco calibration plate 5 corresponds to the 360-degree fixed installation of the line laser scanner fixed support base with a rotating scale.
[0059] ChArUco calibration plates No. 1, No. 2 and No. 3 are composed of the fusion of Zhang's "chessboard" calibration plate and the traditional ArUco calibration plate.
[0060] The host computer is connected to the line laser scanner 1, controls the line laser scanner 1 to rotate according to the scale, and shoots data sets of different angles and different ChArUco calibration plates during the rotation process, and completes subsequent work in the host computer.
[0061] This embodiment also proposes a line scan pose calibration method based on a ChArUco calibration board, the method comprising:
[0062] Obtain the intrinsic reference data of the line laser scanner;
[0063] Collect image data of the ChArUco calibration plate;
[0064] Use the RefineNet network to locate key corner points of image data;
[0065] Based on the positioned image data, combined with the PnP algorithm and the internal reference data of the line laser scanner, several sets of pose values between the coordinate system where the line laser scanner is located and the world coordinate system are obtained, and several sets of pose values are used for iterative optimization to obtain the final pose value.
[0066] Furthermore, the checkerboard and the corner points of the ArUco calibration plate are selected as the key corner points of the image data.
[0067] Furthermore, obtaining the internal reference data of the line laser scanner includes:
[0068] Convert the world coordinate system of the line laser scanner to the camera coordinate system;
[0069] Convert the camera coordinate system to the image coordinate system;
[0070] Convert the image coordinate system to the pixel coordinate system;
[0071] The intrinsic parameter data is calculated by combining the camera coordinate system, image coordinate system and pixel coordinate system.
[0072] Furthermore, using the RefineNet network, the key corner points of the image data are located including:
[0073] The key corner points of the collected ChArUco calibration plate image data set are annotated; the key corner points include the corner points of the preset ArUco calibration plate and the checkerboard calibration plate. The selection of key corner points needs to be symmetrical and cover the preset partial area of the calibration plate;
[0074] The RefineNet network is trained based on an image dataset with key corner points annotated;
[0075] The trained RefineNet network is used to locate key corner points of image data and revise the coordinates of corner points.
[0076] Furthermore, in the RefineNet network, a 4-level cascade structure is formed by 4 RefineNet units, which utilizes multi-level feature extraction and long-distance residual connections to generate high-resolution prediction results.
[0077] This embodiment uses computer vision technology and imaging technology to establish a line laser scanner posture calibration system, and combines the RefineNet neural network and the PnP algorithm to solve the posture relationship between the coordinate system of the line laser scanner and the world coordinate system, and finally obtains accurate posture information through iterative optimization of multiple sets of data.
[0078] Solve the internal parameter matrix of line laser scanner 1, such as Figure 6 As shown, the process of solving the intrinsic parameter matrix of the line laser scanner 1 is actually the process of converting the image from the world coordinate system to the pixel coordinate system. The conversion formula is shown below.
[0079] Convert the world coordinate system to the camera coordinate system:
[0080]
[0081] Convert the camera coordinate system to the image coordinate system:
[0082]
[0083] Convert image coordinates to pixel coordinates:
[0084]
[0085] In summary, combining the above formulas (1)(2)(3)(4), the internal parameter calculation formula of the line laser scanner 1 is obtained as shown below.
[0086]
[0087] Among them, the internal parameters of line laser scanner 1 are:
[0088] After the internal reference calibration of the line laser scanner 1 is completed by Zhang's "chessboard" calibration method, the line laser scanner 1 is controlled by the host computer to rotate at a fixed angle of 6 degrees, and the axis rotates 360 degrees in total. During the rotation process, the ChArUco calibration plate image is continuously captured.
[0089] like Figure 4 As shown, ChArUco calibration plate No. 1 3, ChArUco calibration plate No. 2 4, and ChArUco calibration plate No. 3 5 are all composed of a fusion of a checkerboard calibration plate and an ArUco calibration plate marker, and each ChArUco marker has a specific identifiable ID.
[0090] The line laser scanner 1 starts from the leftmost side (0 degrees) of the No. 1 ChArUco calibration plate 3 and rotates clockwise to shoot 20 groups of images for the No. 1 ChArUco calibration plate 3; the No. 2 ChArUco calibration plate 4 (120-240 degrees) shoots 20 groups of images; the No. 3 ChArUco calibration plate 5 (240-360 degrees) shoots 20 groups of images. Finally, the line laser scanner 1 stops rotating facing the rightmost side of the No. 3 ChArUco calibration plate 5, completing the image data acquisition of the ChAruCo calibration plate.
[0091] For rigid objects, as long as we can repeatedly detect a sufficient number of 3D points in a 2D image, we can use the PnP algorithm to calibrate the pose. The key corner points of the image dataset of the ChArUco calibration plate are annotated, and the following are selected: Figure 5 The key points shown avoid the problem of insufficient feature information of all selected checkerboard corner points or ArUco corner points, which is more conducive to the network to locate key corner points and revise the corner point coordinates. Figure 4 The ChArUco calibration plate described above, in this embodiment, the selection of key corner points includes the corner points of the ArUco calibration plate and the checkerboard calibration plate as much as possible. The selection of key corner points needs to be symmetrical and cover most areas of the calibration plate as much as possible to improve the calibration accuracy. Figure 5 The one shown in the figure is just a selection reference method.
[0092] like Figure 3 As shown in the figure, in the host computer, the RefineNet network locates the key corner points and revises the coordinates of the corner points. The four groups of feature maps on the left, 1 / 32, 1 / 16, 1 / 8, and 1 / 4, correspond to the input of images of different scales. RefineNet uses multi-level feature extraction and long-distance residual connections to generate high-resolution prediction results, and uses 4 RefineNet units to form a 4-level cascade structure, which ultimately locates the key corner points and coordinates better.
[0093] After the RefineNet network locates the key corner points and their coordinates, it combines the PnP algorithm and the relevant parameters of the line laser scanner to solve a set of pose relationships between the coordinate system where the line laser scanner is located and the world coordinate system, as shown in the following formula.
[0094] From the above formula (5), we can know that:
[0095]
[0096] Eliminate z from equation (6) c We can get:
[0097]
[0098]
[0099] Each set of 3D-2D matching points corresponds to two equations, and 6 sets of ChArUco punctuation board key corner points are required, from which the position and posture values of the linear laser scanner 1 can be solved.
[0100]
[0101] like Figure 2 As shown in the figure, after solving a set of pose values, the above operation is repeated to calibrate the pose of 60 sets of ChArUco calibration plates after collection and calibration, and iterative optimization is performed using multiple sets of values to finally obtain the most accurate pose value and complete the overall process.
[0102] This embodiment can realize the full-scale collection of multiple groups of ChArUco calibration plate image data information at different angles, and is less affected by the external environment. The key corner points of the ChArUco calibration plate are located and the corner point coordinates are revised through the RefineNet neural network, and the pose relationship between the coordinate system where the line laser scanner is located and the world coordinate system is solved in combination with the PnP algorithm and the relevant parameters of the line laser scanner. Through the continuous iterative optimization of multiple groups of data, the most accurate pose can be finally obtained, completing the overall pose calibration process of the line laser scanner.
[0103] Compared with the prior art, this embodiment has the following advantages:
[0104] (1) It is less affected by the detection environment and has lower requirements on the external detection environment, so the accuracy of posture calibration will not be greatly affected by the detection environment.
[0105] (2) The RefineNet deep learning network is used to locate the key corner points of the ChArUco calibration plate and revise the coordinates, achieving high positioning accuracy and better detection effect.
[0106] (3) The use of three ChArUco calibration plates avoids the problem that traditional pose calibration is limited to a single plane, and has a wider range of application scenarios.
[0107] (4) A multi-group optimization method is used to uniformly iteratively optimize the pose values obtained from multiple groups to obtain the final accurate pose value, thereby reducing the possible errors in single data.
[0108] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A line scan pose calibration system based on ChArUco calibration board, characterized in that: include: Line laser scanner, connection device, several ChArUco calibration plates, telescopic device, rotatable support device; The telescopic device is fixed on the rotatable supporting device, and the line laser scanner is fixed on the other end of the telescopic device; The ChArUco calibration plates are centered on the rotatable support device and are evenly distributed at preset angles between each other; The line laser scanner collects image data of the ChArUco calibration plate through the rotation of the rotatable support device, and uploads the image data to the host computer through the connecting device for position and posture marking.
2. The line scan pose calibration system based on the ChArUco calibration board according to claim 1, characterized in that: The ChArUco calibration plate includes: a first ChArUco calibration plate, a second ChArUco calibration plate, and a third ChArUco calibration plate; The first ChArUco calibration plate is fixedly installed corresponding to 120 degrees of the rotatable support device; the second ChArUco calibration plate is fixedly installed corresponding to 240 degrees of the rotatable support device; and the third ChArUco calibration plate is fixedly installed corresponding to 360 degrees of the rotatable support device.
3. The line scan pose calibration system based on the ChArUco calibration board according to claim 1, characterized in that: The ChArUco calibration plate is composed of a fusion of a checkerboard calibration plate and an ArUco calibration plate.
4. A line scan pose calibration method based on ChArUco calibration board, characterized in that: Used to implement the line scan pose calibration system based on the ChArUco calibration board as described in any one of claims 1 to 3, the method comprising: Obtain the intrinsic reference data of the line laser scanner; Collect image data of the ChArUco calibration plate; Using the RefineNet network, key corner points of the image data are located; Based on the positioned image data, combined with the PnP algorithm and the internal reference data of the line laser scanner, several sets of pose values between the coordinate system where the line laser scanner is located and the world coordinate system are obtained, and several sets of pose values are used for iterative optimization to obtain the final pose value.
5. The line scan pose calibration method based on the ChArUco calibration board according to claim 4 is characterized in that: The checkerboard and the corner points of the ArUco calibration plate are selected as the key corner points of the image data.
6. The line scan pose calibration method based on the ChArUco calibration board according to claim 4, characterized in that: The internal parameter data of the line laser scanner include: Converting the world coordinate system of the line laser scanner into a camera coordinate system; Convert the camera coordinate system into an image coordinate system; Converting the image coordinate system into a pixel coordinate system; The intrinsic parameter data is calculated by combining the camera coordinate system, the image coordinate system and the pixel coordinate system.
7. The line scan pose calibration method based on the ChArUco calibration board according to claim 4, characterized in that: Using the RefineNet network, locating key corner points of the image data includes: Annotate the key corner points of the collected ChArUco calibration plate image dataset; The RefineNet network is trained based on an image dataset with key corner points annotated; The trained RefineNet network is used to locate key corner points of the image data and revise the coordinates of the corner points.
8. The line scan pose calibration method based on the ChArUco calibration board according to claim 7, characterized in that: In the RefineNet network, four RefineNet units form a four-level cascade structure, and multi-level feature extraction and long-distance residual connections are used to generate high-resolution prediction results.
9. The line scan pose calibration method based on the ChArUco calibration board according to claim 7, characterized in that: The set of pose values is: Among them, R represents the rotation matrix, T represents the translation matrix, and f ab Represents the calibration result parameters, a=1, 2, 3, b=1, 2, 3, 4.
Citation Information
Patent Citations
Full-view linear laser scanning 3D imaging calibration device and full-view linear laser scanning 3D imaging calibration method
CN106056587A
Line laser scanning three-dimensional measurement calibration method
CN111272102A
Space-time calibration method for self-rotation line structured light scanning system
CN112833791A
Image space projection splicing method based on integrated calibration parameters
CN112927133A
Deep neural network pose estimation system
CN113272821A
