A motion capture scanning joint calibration and pose solving method based on an L-shaped positioning rod

Abstract

The application discloses a motion capture scanning combined calibration and pose solution method based on an L-shaped positioning rod, belongs to the technical field of motion capture and three-dimensional scanning combined measurement, and carries out overall three-dimensional scanning on the L-shaped positioning rod and a rigid body to be measured in a static state, constructs an L-shaped positioning rod coordinate system and a real target coordinate system of the rigid body to be measured in a three-dimensional scanning model coordinate system respectively, solves the relative pose matrix of both, and corrects the motion capture rigid body to be measured coordinate system based on the same, combines the pose of the L-shaped positioning rod in a motion capture room coordinate system, and obtains the real-time pose of the real target coordinate system of the rigid body to be measured; a marker point is only used for rigid body tracking and state checking and does not participate in the definition of the real target coordinate system. The application considers the scanning geometric accuracy and the motion capture dynamic tracking capability, and can be applied to robot calibration, optical adjustment, semi-physical simulation and other measurement scenes.

Description

Technical Field

[0001] This invention belongs to the field of motion capture and 3D scanning joint measurement technology, specifically involving a motion capture and scanning joint calibration and pose calculation method based on an L-shaped positioning rod. Background Technology

[0002] Scenarios such as robot calibration, optical assembly and adjustment, semi-physical simulation of spatial targets, and multi-rigid-body assembly measurement place high demands on the real-time performance and accuracy of target rigid body pose measurement. 3D scanning systems can acquire high-precision static 3D geometric models of the object under test, extracting key geometric references such as the centroid, optical center, and assembly reference points; motion capture systems can achieve real-time dynamic tracking of rigid bodies using marker points. However, the two measurement systems operate in independent coordinate systems, and there is a lack of a stable and unified common reference between the 3D scanning model coordinate system and the motion capture room coordinate system. This makes it difficult to directly map the high-precision geometric definitions obtained from 3D scanning into the real-time motion capture measurement results, and the static geometric information cannot serve the real-time pose calculation during the object's motion.

[0003] When a motion capture system establishes a rigid body, it typically generates a rigid body coordinate system automatically based on the spatial distribution of marker points. This default coordinate system only reflects the geometric distribution of the marker point combination and may not necessarily coincide with the actual target coordinate system of the rigid body being measured. For objects such as cubic mirrors, optical lens groups, and robotic arm end effectors, engineering practice often focuses on the pose of functional references such as the center of mass, optical center, and tool center point. However, the coordinate system generated by the default motion capture system introduces both positional and attitude deviations. If marker points are placed around the actual target point to reduce these deviations, the system is limited by structural space, reflections, and motion occlusion, which can easily lead to marker loss and rigid body tracking failure. Existing technologies also employ manual offset compensation, dedicated calibration fixtures, or external equipment such as laser trackers to register two systems. However, these methods generally suffer from drawbacks such as large human errors, modification of the object structure, high equipment costs, and cumbersome operation procedures.

[0004] In multi-rigid-body collaborative measurement scenarios, the aforementioned problems are further amplified. If each rigid body directly uses the default coordinate system output pose obtained by fitting marker points, the calculated relative poses between the multiple rigid bodies do not represent the actual relative relationships between the real target coordinate systems of each measured rigid body. This error will propagate down to subsequent calibration and assembly calculation stages, reducing the reliability of the experimental results. Therefore, a joint calibration method that balances ease of operation and measurement accuracy is urgently needed to achieve real-time pose output of the real target coordinate systems of one or more measured rigid bodies. Summary of the Invention

[0005] To address the aforementioned problems in the prior art, the present invention is achieved through the following technical solution: This invention provides a method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod, characterized by the following steps: The L-shaped positioning rod in the motion capture system and the rigid body to be measured are placed statically in the same experimental space for three-dimensional scanning to obtain a three-dimensional model, which is then imported into three-dimensional model processing software. In the coordinate system of the 3D scanning model, the coordinate system of the L-shaped positioning rod is established based on the marker ball on the L-shaped positioning rod, and the pose matrix of the L-shaped positioning rod coordinate system is obtained based on the coordinate system of the 3D scanning model. In the coordinate system of the three-dimensional scanning model, the geometric reference features of the rigid body to be tested are extracted. Based on the geometric reference features, the real target coordinate system of the rigid body to be tested is established. The pose matrix of the real target coordinate system is obtained based on the coordinate system of the three-dimensional scanning model. Based on the pose matrix of the L-shaped positioning rod coordinate system and the pose matrix of the real target coordinate system, the relative pose matrix of the real target coordinate system of the rigid body under test relative to the L-shaped positioning rod coordinate system is obtained through matrix operations. In the motion capture software, the pose of the L-shaped positioning rod coordinate system in the motion capture room coordinate system is obtained. Based on the relative pose matrix, the pose of the real target coordinate system of the rigid body under test in the motion capture room coordinate system is obtained.

[0006] Before the 3D scanning step, the following steps are also included: The motion capture system identifies the marker ball attached to the L-shaped positioning rod and performs field calibration to establish the coordinate system of the motion capture room; after the field calibration is completed, the physical position of the L-shaped positioning rod remains fixed.

[0007] The pose matrix of the L-shaped positioning rod coordinate system is obtained from the coordinate system of the 3D scanning model, including: Obtain the centroids of two marker spheres along the major axis and two along the minor axis of the L-shaped positioning rod. Define the line connecting the two centroids along the major axis as the X-axis of the L-shaped positioning rod coordinate system; define the line connecting the two centroids along the minor axis as the Y-axis. According to the right-hand rule, the cross product of the X-axis and Y-axis directions yields the normal vector of the reference plane, which is then used as the Z-axis. Shifting the Z-axis by a preset offset distance yields the origin of the L-shaped positioning rod coordinate system. Based on the coordinate system S of the 3D scanning model and the origin of the L-shaped positioning rod coordinate system, obtain the pose matrix of the L-shaped positioning rod coordinate system. .

[0008] Obtain the centroids of the two marker spheres along the major axis and the two marker spheres along the minor axis of the L-shaped positioning rod, including: Within the 3D scanning model, select the spherical point cloud region corresponding to each marker sphere of the L-shaped positioning rod, perform geometric solution on the selected spherical point cloud, and obtain the centroid of each marker sphere.

[0009] In the coordinate system of the 3D scanning model, the geometric reference features of the rigid body to be measured are extracted. Based on the geometric reference features, the true target coordinate system of the rigid body to be measured is established, including: In the 3D scanning model, select at least two mutually perpendicular reference plane point cloud regions for the rigid body to be measured, and perform plane fitting on the point clouds of each plane to obtain the normal vectors corresponding to each reference plane; determine the coordinate axis directions of the temporary coordinate system based on the reference plane normal vectors, and then translate the temporary coordinate system according to the design dimensions or assembly offset of the rigid body to be measured to obtain the real target coordinate system of the rigid body to be measured.

[0010] Based on the pose matrix of the L-shaped positioning rod coordinate system and the pose matrix of the real target coordinate system, the relative pose matrix of the rigid body under test relative to the L-shaped positioning rod coordinate system is obtained through matrix operations. The formula is as follows:

[0011] in, Let be the pose matrix of the L-shaped positioning rod coordinate system. The pose matrix is ​​the coordinate system of the real target.

[0012] After obtaining the relative pose matrix, the following steps are also included: In motion capture software, several marker points are selected on the rigid body to be measured, and a coordinate system of the rigid body to be measured is established according to the spatial distribution of the marker points. The coordinate system of the rigid body to be measured is used for the identification and monitoring of marker point occlusion or detachment.

[0013] The poses of the L-shaped positioning rod coordinate system and the actual target coordinate system of the rigid body under test in the 3D scanning model coordinate system are both expressed using a 4th-order homogeneous transformation matrix, with the following expression:

[0014] Wherein, RSA is a 3×3 rotation matrix. The position vector of the 3D scanning model in coordinate system S is the origin of the coordinate system.

[0015] The rigid body to be tested is a cubic mirror.

[0016] When there are multiple rigid bodies to be measured, the steps of 3D scanning, extracting geometric reference features and establishing the corresponding real target coordinate system are performed for each rigid body to be measured. The relative pose matrix of the real target coordinate system of each rigid body to be measured relative to the L-shaped positioning rod coordinate system is solved. Combined with the pose of the L-shaped positioning rod coordinate system in the motion capture room coordinate system, the pose of the real target coordinate system of each rigid body to be measured in the motion capture room coordinate system is calculated. Based on the pose of the real target coordinate system of each rigid body to be measured, the relative pose between the real target coordinate systems of multiple rigid bodies to be measured is calculated.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention reuses the existing L-shaped positioning rod of the motion capture system as the common reference for the two systems. There is no need to configure additional calibration plates, laser trackers and other external equipment. Through a single overall three-dimensional scan, the geometric definition of the target in the scanned model can be directly associated with the coordinate system of the motion capture room. The operation is simple and the experimental hardware cost is reduced. 2. Define the real target coordinate system of the rigid body to be measured independently in the 3D scanning model. The origin of the coordinate system can be set at functional positions such as the centroid, optical center, tool center point, assembly reference point, etc., and is decoupled from the marker point layout. The marker is only used for rigid body recognition and tracking and can be placed in areas that are not easily obstructed to avoid rigid body tracking failure caused by the spatial and reflective limitations of the target body and to give full play to the geometric measurement accuracy of 3D scanning. 3. The coordinate system of the motion capture rigid body is corrected by the relative pose matrix, and the real-time pose of the target coordinate system is directly output. Multiple rigid bodies can share the same set of reference and calculation logic to complete the calculation of the relative pose between each target in the coordinate system. It is compatible with a variety of measurement objects such as cubic mirrors, robotic arm end tools, and precision assemblies, and has a wide range of applications.

[0018] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] Figure 1 This is a flowchart of a motion capture scanning joint calibration and pose calculation method based on an L-shaped positioning rod provided by the present invention; Figure 2 This is a schematic diagram of the spherical point cloud region corresponding to the marker sphere selection provided by the present invention; Figure 3 This is a complete schematic diagram of the L-shaped positioning rod provided by the present invention; Figure 4 This is a schematic diagram of the L-shaped positioning rod coordinate system provided by the present invention; Figure 5 This is a schematic diagram of the selection of point cloud in a cubic mirror plane provided by the present invention; Figure 6This is a schematic diagram of the true target coordinate system of the cubic mirror provided by the present invention; Figure 7 This is a schematic diagram of arranging marker points around a cubic mirror provided by the present invention; Figure 8 This is a schematic diagram of modifying the pose of a cubic mirror rigid body in a motion capture system, provided by the present invention. Detailed Implementation

[0020] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following detailed description of the solution according to the present invention is provided in conjunction with the accompanying drawings and specific embodiments.

[0021] The foregoing and other technical contents, features, and effects of the present invention will be clearly presented in the following detailed description of specific embodiments in conjunction with the accompanying drawings. Through the description of the specific embodiments, a more in-depth and concrete understanding can be gained of the technical means and effects adopted by the present invention to achieve its intended purpose. However, the accompanying drawings are for reference and illustration only and are not intended to limit the technical solutions of the present invention.

[0022] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that an article or apparatus comprising a list of elements includes not only those elements but also other elements not expressly listed. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or apparatus that includes said element.

[0023] like Figure 1 The diagram shows a flowchart of a motion capture scanning joint calibration and pose calculation method based on an L-shaped positioning rod provided by the present invention, which includes the following steps: 1. Preliminary preparation stage: The L-shaped positioning rod in the motion capture system and the rigid body to be measured are placed statically in the same experimental space for three-dimensional scanning to obtain a three-dimensional model, which is then imported into three-dimensional model processing software.

[0024] Specifically, in the initial static state, the L-shaped positioning rod of the motion capture system and the rigid body under test are placed in the same experimental space. Several marker points are attached to the rigid body under test or its support. The L-shaped positioning rod, the rigid body under test, the marker points, and the mounting bracket are all kept stationary. The rigid body under test is mounted on the mounting bracket. The marker points are only used for motion capture identification and tracking of the rigid body. They do not need to be symmetrically arranged around the real target point, nor is it required that the origin of the coordinate system of the motion capture rigid body under test obtained by the markers coincide with the real target point. They can be placed in positions where occlusion is unlikely. The beneficial effect of this step is that the marker placement is not limited by the position of the real target point, reducing the problem of occlusion and loss of points, and improving the stability of motion capture tracking.

[0025] Next, a 3D scan is performed on the stationary L-shaped positioning rod and the rigid body under test to obtain a 3D model containing their geometric shapes. This model is then imported into Blender 3D model processing software for subsequent coordinate system establishment and pose calculation. Before the coordinate system is established and the relative pose matrix is ​​solved, the L-shaped positioning rod and the rigid body under test must not be displaced to avoid misalignment between the model's geometric relationship and physical space. This step, through a single overall scan, incorporates the L-shaped positioning rod and the rigid body under test, which serve as a common reference, into the same model, providing a reliable foundation for the coordinate unification of the two measurement systems and ensuring the accuracy of pose calculation.

[0026] 2. Offline modeling and calibration stage: 2.1 Selection of L-shaped positioning rod features and establishment of L-shaped positioning rod coordinate system: In the coordinate system of the 3D scanning model, the coordinate system of the L-shaped positioning rod is established based on the marker ball on the L-shaped positioning rod, and the pose matrix of the L-shaped positioning rod coordinate system is obtained based on the coordinate system of the 3D scanning model.

[0027] Specifically, in the 3D scanning model corresponding to the coordinate system S of the 3D scanning model, the L-shaped positioning rods used by the motion capture system are identified. For each marker sphere of the L-shaped positioning rod, its corresponding spherical point cloud region is selected; for example... Figure 2 The diagram shows a schematic of the selected spherical point cloud region corresponding to the marker sphere. For each marker sphere, instead of directly selecting a single discrete point within the 3D scan model as the sphere center, a region of point cloud on the surface of the marker sphere is selected. Geometric solutions are then performed on the selected spherical point cloud to obtain the centroid coordinates of the corresponding marker sphere. This processing method can reduce the adverse effects of scanning noise, missing point clouds, and manual point selection errors on the centroid extraction results of the marker sphere.

[0028] Let the centroids of the two marker spheres along the major axis of the L-shaped positioning rod be P1 and P2, and the centroids of the two marker spheres along the minor axis be P3 and P4. The order in which the marker spheres are selected must be consistent with the order in which the motion capture system uses the L-shaped positioning rod to define the coordinate system of the motion capture room. Subsequent coordinate axis directions, origin offset directions, and right-handed coordinate system rules must all be configured according to the motion capture system's rules for establishing the L-shaped positioning rod coordinate system, ensuring that the L-shaped positioning rod coordinate system established in the 3D scanning model has a consistent spatial definition with the motion capture room coordinate system. Figure 3 The image shows a complete schematic diagram of the L-shaped positioning rod.

[0029] Based on the coordinates of the four marker sphere centroids in the 3D scanning model coordinate system S, an L-shaped positioning rod coordinate system is established. Specifically, the line connecting the two marker sphere centroids P1 and P2 along the major axis is defined as the X-axis of the L-shaped positioning rod coordinate system; the line connecting the two marker sphere centroids P3 and P4 along the minor axis is defined as the Y-axis of the L-shaped positioning rod coordinate system. The lines connecting the major and minor axes correspond to the extension directions of the two arms of the L-shaped positioning rod, respectively. The reference plane where the L-shaped positioning rod is located is determined by the two axes, and the intersection of the two axes is used as the reference point before the origin offset.

[0030] After determining the X-axis and Y-axis directions, according to the right-hand coordinate system rule, perform a cross product operation between the X-axis and Y-axis vectors to obtain the normal vector of the reference plane, which is used as the Z-axis direction of the L-shaped positioning rod coordinate system. For example... Figure 4 The diagram shows the coordinate system of the L-shaped positioning rod. In actual operation, the origin of the L-shaped positioning rod in the motion capture system is not located on the plane formed by the centroids of the four marker spheres, but rather at a position offset by a certain distance along the -Z direction. Therefore, in the 3D scanning model, the intersection of the major and minor axes is used as the reference point, and a preset offset distance is moved along the -Z direction. In this embodiment, the preset distance is 25mm, resulting in the origin of the L-shaped positioning rod coordinate system consistent with the system establishment rules of the motion capture system.

[0031] In the 3D scanning model coordinate system S, the centroids of the four marker spheres each possess spatial coordinates in S. Based on this set of coordinates, the coordinate axis directions of the L-shaped positioning rod coordinate system and the position of the origin in S are calculated. The coordinate axis directions are assembled into a rotation matrix, and the origin coordinates are used as a translation vector. These are combined to obtain the pose matrix of the L-shaped positioning rod coordinate system in the 3D scanning model coordinate system S, denoted as... The pose matrix represents the position and orientation of the L-shaped positioning rod coordinate system in the three-dimensional scanning model coordinate system S.

[0032] Since the coordinate system of the motion capture room and the L-shaped positioning rod have a definite correspondence, and the coordinate system is established based on the same physical L-shaped positioning rod in the 3D scanning model; when the marker ball selection numbering order, X-axis direction, Y-axis direction, Z-axis direction and the preset offset distance of the origin along the -Z direction are all consistent with the motion capture system, the geometric relationship in the 3D scanning model can be uniformly mapped to the coordinate system of the motion capture room by relying on the L-shaped positioning rod.

[0033] 2.2 Selection of planar points for the cubic mirror and establishment of the cubic mirror coordinate system: In the coordinate system of the 3D scanning model, the geometric reference features of the rigid body to be measured are extracted. Based on the geometric reference features, the true target coordinate system of the rigid body to be measured is established, and the pose matrix of the true target coordinate system is obtained based on the coordinate system of the 3D scanning model. The rigid body to be measured is preferably a cubic mirror.

[0034] In the 3D scanning model, two mutually perpendicular reference planes are selected for the rigid body under test. A point cloud region is selected on each reference plane to fit the spatial equation of the corresponding plane. When selecting the point cloud region, areas with broken edges, high scanning noise, severe reflection, or occlusion should be avoided. Priority should be given to flat, complete regions that can represent the true orientation of the plane. For example... Figure 5 The diagram shown is a schematic of the selection of point clouds in the plane of a cubic mirror.

[0035] Plane fitting is performed on the first reference plane point cloud region to obtain the normal vector corresponding to the reference plane; plane fitting is performed on the second reference plane point cloud region to obtain the normal vector corresponding to the reference plane; the normal vectors corresponding to the two mutually perpendicular reference planes of the rigid body under test are used to determine the directions of the two coordinate axes of the real target coordinate system of the rigid body under test.

[0036] After obtaining the normal vectors corresponding to the two reference planes, select the center point of one of the reference planes or a reference point that can characterize the center position of that plane. The benchmark point The origin of the temporary coordinate system is used as the reference plane. The directions of the two coordinate axes of the temporary coordinate system are determined by the normal vectors of the two reference planes. The direction of the remaining third coordinate axis is obtained by the cross product of the two normal vectors. The positive directions of each coordinate axis can be uniformly set according to the design structure of the rigid body to be measured, the orientation of the working surface, the installation direction, or the coordinate definition agreed upon in advance in the experiment.

[0037] Specifically, when the real target point is the centroid of the cubic mirror, and the point... When the center position of a certain outer surface of the cube mirror is corresponding to the position of the temporary coordinate system, the temporary coordinate system is translated into the cube mirror along the inner normal direction of the plane by half a side length, so that the origin of the coordinate system moves to the centroid of the cube mirror. If the outer side of the cube mirror has a mounting base, adhesive layer, clamp or other additional structure, the translation amount is no longer directly taken as half a side length, and it is necessary to make translation transformation correction in combination with the cube mirror design size, assembly relationship or the offset amount obtained by actual measurement.

[0038] The coordinate system obtained after translation transformation is the true target coordinate system of the rigid body under test. ;like Figure 6 The diagram shows the true target coordinate system of the cubic mirror. The orientation of this true target coordinate system is determined by the normal vector of the reference plane of the rigid body under test. The origin is determined by the center point of the reference plane combined with known geometric dimensions or measured offsets. Its definition does not depend on the spatial distribution of marker points, but is jointly determined by the 3D scanning model and the structure of the rigid body under test.

[0039] In the 3D scanning model coordinate system S, the origin and three coordinate axes of the real target coordinate system of the rigid body under test can be obtained as spatial positions and orientation vectors in S. Based on the above information, a 3×3 rotation matrix is ​​assembled from the coordinate axes of the real target coordinate system. The spatial coordinates of the origin of the real target coordinate system in the 3D scanning model coordinate system S are used as translation vectors. The combination of the rotation matrix and the translation vectors yields the pose matrix of the real target coordinate system in the 3D scanning model coordinate system S, denoted as […]. The pose matrix represents the position and orientation of the real target coordinate system of the rigid body under test in the three-dimensional scanning model coordinate system S.

[0040] 2.3 Solving the relative pose of a rigid body and a cube mirror with respect to an L-shaped positioning rod in Blender: In Blender or other 3D modeling software, read the coordinate system of the L-shaped positioning rod. And the true target coordinate system of the cube mirror In the coordinate system of the 3D scanning model The position and orientation of the L-shaped positioning rod in the coordinate system are as follows: The pose of the cubic mirror in the true target coordinate system is denoted as .

[0041] Based on the pose matrix of the L-shaped positioning rod coordinate system pose matrix in the true target coordinate system The relative pose matrix of the rigid body under test relative to the L-shaped positioning rod coordinate system is obtained through matrix operations. The formula is as follows:

[0042] in, Let be the pose matrix of the L-shaped positioning rod coordinate system. The pose matrix is ​​the coordinate system of the real target.

[0043] This matrix records the actual spatial relationship of the cubic mirror's centroid, optical center, or other target points under the L-shaped positioning rod reference at the moment of scanning, and is a core parameter for subsequent correction of the motion capture rigid body coordinate system. Since the L-shaped positioning rod is used in the motion capture system to establish or associate the room coordinate system... It can be further used to map the coordinate system of the cubic mirror real target determined in the scanning model to the coordinate system of the motion capture room.

[0044] The poses of the L-shaped positioning rod coordinate system and the actual target coordinate system of the rigid body under test in the Blender 3D scanning model coordinate system are both expressed using a 4th-order homogeneous transformation matrix, as follows:

[0045] Among them, It is a 3×3 rotation matrix. The position vector of the 3D scanning model in coordinate system S is the origin of the coordinate system.

[0046] The Euler rotation angles of the L-shaped positioning rod coordinate system and the actual target coordinate system of the rigid body under test are read from the Blender interface. The unit of the angle is degrees. Before performing matrix operations, the angles need to be converted to radians before generating the corresponding 3×3 rotation matrix.

[0047] To convert Euler angles to rotation vectors, the following formula is required:

[0048] in, , and These are rotation matrices around the X-axis, rotation matrices around the Y-axis, and rotation matrices around the Z-axis, respectively. , and These are the Euler rotation angles about the X-axis, Y-axis, and Z-axis, respectively.

[0049] The three basic rotation matrices are as follows:

[0050] 3. Motion capture correction and online output stage: In the motion capture software, the pose of the L-shaped positioning rod coordinate system in the motion capture room coordinate system is obtained. Based on the relative pose matrix, the pose of the real target coordinate system of the rigid body under test in the motion capture room coordinate system is obtained.

[0051] like Figure 7The diagram shows a schematic of marker points arranged around a cubic mirror. In the motion capture software, several marker points are selected on the rigid body to be measured or its support to establish a motion-captured rigid body, thereby generating a coordinate system for the motion-captured rigid body. This coordinate system is automatically generated by fitting the spatial distribution of the marker points, and its origin and coordinate axes may not necessarily coincide with the actual target coordinate system of the rigid body. This coordinate system is only used to complete the rigid body binding configuration of the marker points on the rigid body to be measured, to realize the marker point recognition status verification, to determine whether the marker points are occluded or detached, and only represents the overall motion state of the marker points. The pose originally output by the coordinate system of the motion-captured rigid body is not used as input for the pose calculation of this invention, but only as reference data for accuracy comparison. This invention does not rely on the original pose of the coordinate system of the motion-captured rigid body to solve for the actual target coordinate system of the rigid body to be measured, nor does it directly use it as the final actual target coordinate system output by this method.

[0052] After the motion capture system completes the rigid body establishment, the origin position and coordinate axis direction of the rigid body to be captured are corrected based on the obtained pose relationship between the rigid body to be captured and the L-shaped positioning rod. The relative pose matrix of the real target coordinate system of the rigid body to be captured relative to the L-shaped positioning rod coordinate system, obtained in the coordinate system of the 3D scanning model, is written into the motion capture software to align the rigid body coordinate system in the motion capture software with the real target coordinate system of the rigid body to be captured.

[0053] like Figure 8 The diagram illustrates the modification of the rigid body pose of the cube mirror in the motion capture system. After this modification, the marker point still performs the functions of rigid body recognition and real-time tracking. The origin and coordinate axis directions of the rigid body coordinate system are determined by the actual target coordinate system of the rigid body under test, obtained from the 3D scanning model. The pose of the L-shaped positioning rod coordinate system in the motion capture room coordinate system is obtained in the motion capture software. Combined with the aforementioned relative pose matrix, the pose of the actual target coordinate system of the rigid body under test in the motion capture room coordinate system can be obtained. At this point, the rigid body pose output by the motion capture system is no longer the pose corresponding to the default geometric center of the marker point, but rather the pose corresponding to the centroid of the rigid body under test or other preset actual target points.

[0054] 4. Multi-rigid-body pose calculation: For multiple rigid bodies to be measured, their respective true target coordinate systems can be established by following the steps described above. , … The system calculates the relative pose of each target coordinate system with respect to the L-shaped positioning rod coordinate system. The motion capture system establishes a corresponding marker rigid body for each target rigid body. , … Then, the corresponding fixed correction relationships are written or saved respectively. Multiple real target coordinate systems are simultaneously output in the motion capture room coordinate system. The real-time pose of the target. The relative pose between any two real target coordinate systems can be calculated by the following formula:

[0055] in, This represents the coordinate system of the actual target of the rigid body under test at time t. Relative to the true target coordinate system of the rigid body under test The 4th order homogeneous relative pose matrix; This represents the coordinate system of the actual target of the rigid body under test at time t. The 4th-order homogeneous relative pose matrix relative to the reference coordinate system R (motion capture room coordinate system); This represents the coordinate system of the actual target of the rigid body under test at time t. The fourth-order homogeneous relative pose matrix relative to the reference coordinate system R (motion capture room coordinate system).

[0056] The results represent the relative position and orientation between real target coordinate systems, and can be used in scenarios such as cube mirror positioning, robotic arm end-effector calibration, optical device attitude measurement, and multi-rigid-body assembly relationship calculation.

[0057] The scanning phase of this invention is used to determine the position of the real target point and the direction of the target coordinate axis, while the motion capture phase is used to track the target's movement over time in real time. An L-shaped positioning rod serves as a common reference between the two systems, enabling the geometric definitions in the scanning model to be mapped to the motion capture room coordinate system. The marker point only undertakes the real-time tracking task and no longer determines the position and direction of the real target coordinate system. This reduces the difficulty of marker point placement, decreases the risk of rigid body loss due to occlusion, and improves the accuracy of the real target pose output, including the centroid of the cubic mirror, the optical center, and the tool center point.

[0058] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod, characterized in that, Includes the following steps: The L-shaped positioning rod in the motion capture system and the rigid body to be measured are placed statically in the same experimental space for three-dimensional scanning to obtain a three-dimensional model, which is then imported into three-dimensional model processing software. In the coordinate system of the 3D scanning model, the coordinate system of the L-shaped positioning rod is established based on the marker ball on the L-shaped positioning rod, and the pose matrix of the L-shaped positioning rod coordinate system is obtained based on the coordinate system of the 3D scanning model. In the coordinate system of the three-dimensional scanning model, the geometric reference features of the rigid body to be tested are extracted. Based on the geometric reference features, the real target coordinate system of the rigid body to be tested is established. The pose matrix of the real target coordinate system is obtained based on the coordinate system of the three-dimensional scanning model. Based on the pose matrix of the L-shaped positioning rod coordinate system and the pose matrix of the real target coordinate system, the relative pose matrix of the real target coordinate system of the rigid body under test relative to the L-shaped positioning rod coordinate system is obtained through matrix operations. In the motion capture software, the pose of the L-shaped positioning rod coordinate system in the motion capture room coordinate system is obtained. Based on the relative pose matrix, the pose of the real target coordinate system of the rigid body under test in the motion capture room coordinate system is obtained.

2. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 1, characterized in that, Before the three-dimensional scanning step, the following steps are also included: The motion capture system identifies the marker ball attached to the L-shaped positioning rod and performs field calibration to establish the coordinate system of the motion capture room; after the field calibration is completed, the physical position of the L-shaped positioning rod remains fixed.

3. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 1, characterized in that, The pose matrix of the L-shaped positioning rod coordinate system obtained from the coordinate system of the three-dimensional scanning model includes: Obtain the centroids of two marker spheres along the major axis and two along the minor axis of the L-shaped positioning rod. Define the direction of the line connecting the centroids of the two marker spheres along the major axis as the X-axis of the L-shaped positioning rod coordinate system; define the direction of the line connecting the centroids of the two marker spheres along the minor axis as the Y-axis of the L-shaped positioning rod coordinate system. According to the right-hand coordinate system rule, the normal vector of the reference plane obtained by the cross product of the X-axis and Y-axis directions is used as the Z-axis direction. Move the Z-axis direction by a preset offset distance to obtain the origin of the L-shaped positioning rod coordinate system. Based on the coordinate system S of the 3D scanning model and the origin of the L-shaped positioning rod coordinate system, obtain the pose matrix of the L-shaped positioning rod coordinate system.

4. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 3, characterized in that, The process of obtaining the centroids of the two marker spheres along the major axis and the two centroids of the marker spheres along the minor axis of the L-shaped positioning rod includes: Within the 3D scanning model, select the spherical point cloud region corresponding to each marker sphere of the L-shaped positioning rod, perform geometric solution on the selected spherical point cloud, and obtain the centroid of each marker sphere.

5. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 1, characterized in that, The step of extracting the geometric reference features of the rigid body under test in the coordinate system of the three-dimensional scanning model, and establishing the real target coordinate system of the rigid body under test based on the geometric reference features, includes: In the 3D scanning model, select at least two mutually perpendicular reference plane point cloud regions for the rigid body to be measured, and perform plane fitting on the point clouds of each plane to obtain the normal vectors corresponding to each reference plane; determine the coordinate axis directions of the temporary coordinate system based on the reference plane normal vectors, and then translate the temporary coordinate system according to the design dimensions or assembly offset of the rigid body to be measured to obtain the real target coordinate system of the rigid body to be measured.

6. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 1, characterized in that, The relative pose matrix of the rigid body under test relative to the L-shaped positioning rod coordinate system is obtained through matrix operations based on the pose matrix of the L-shaped positioning rod coordinate system and the pose matrix of the real target coordinate system. The formula is as follows: in, Let be the pose matrix of the L-shaped positioning rod coordinate system. The pose matrix is ​​the coordinate system of the real target.

7. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 1, characterized in that, After obtaining the relative pose matrix, the process further includes: In motion capture software, several marker points are selected on the rigid body to be measured, and a coordinate system of the rigid body to be measured is established according to the spatial distribution of the marker points. The coordinate system of the rigid body to be measured is used for the identification and monitoring of marker point occlusion or detachment.

8. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 1, characterized in that, The poses of the L-shaped positioning rod coordinate system and the actual target coordinate system of the rigid body under test in the 3D scanning model coordinate system are both expressed using a 4th-order homogeneous transformation matrix, with the following expression: Wherein, RSA is a 3×3 rotation matrix. The position vector of the 3D scanning model in coordinate system S is the origin of the coordinate system.

9. The method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod according to claim 1, characterized in that, The rigid body to be tested is a cubic mirror.

10. A method for joint calibration and pose calculation of motion capture scanning based on an L-shaped positioning rod, characterized in that, When there are multiple rigid bodies to be measured, the steps of 3D scanning, extracting geometric reference features and establishing the corresponding real target coordinate system are performed for each rigid body to be measured. The relative pose matrix of the real target coordinate system of each rigid body to be measured relative to the L-shaped positioning rod coordinate system is solved. Combined with the pose of the L-shaped positioning rod coordinate system in the motion capture room coordinate system, the pose of the real target coordinate system of each rigid body to be measured in the motion capture room coordinate system is calculated. Based on the pose of the real target coordinate system of each rigid body to be measured, the relative pose between the real target coordinate systems of multiple rigid bodies to be measured is calculated.

Images