Calculation method for moxibustion physiotherapy position

Abstract

The invention relates to the technical field of machine vision, solves the technical problem of deviation of an actual position caused by posture change in a moxibustion process, and particularly relates to a calculation method for a moxibustion physiotherapy position, which comprises the following steps of: calibrating acquisition equipment at the tail end of a moxibustion robot and acquiring data; establishing a corresponding matching coordinate system on the body point cloud data and the fractal digital meridian graph; performing preliminary registration to obtain a conversion equation between the two; determining a correction scalar according to the conversion equation; carrying out pose correction to obtain an actual position; the actual position is fed back to the moxibustion robot system. According to the method, the corrected actual position coordinates are fed back to the moxibustion robot system, so that the moxibustion robot can still quickly position the corresponding acupuncture points when the posture of the user changes, and the corresponding acupuncture points can be positioned without a deviation phenomenon when the moxibustion robot executes techniques such as positioning, pecking, rotating and transporting in the moxibustion process; and meanwhile, the corresponding deviation is eliminated by automatically adapting to the change of the pose.

Description

Technical Field

[0001] This invention relates to the field of machine vision technology, and in particular to a calculation method for moxibustion therapy locations. Background Technology

[0002] With the increasing demand for health and the continuous development of technology, self-service moxibustion robots are expected to further expand their application areas in the future. These robots utilize the heat generated by burning moxa to accurately deliver heat to acupoints through a specialized robotic arm and control system. This achieves the traditional Chinese medicine effects of unblocking meridians, harmonizing qi and blood, and strengthening the body's resistance to pathogens. Furthermore, the robot maintains the burning temperature of the moxa stick within a suitable range, avoiding the risk of burns and providing continuous and stable heat stimulation.

[0003] A search revealed that CN110037915A, an invention titled "A Self-Service Moxibustion Therapy Robot and Its Control Method," allows for automated moxibustion through programmed settings. Following a pre-defined coordinate system and the desired sequence of acupoints, the robot performs moxibustion sequentially, fully automatically without human intervention. Another example is CN114939060A, an invention titled "A Control Method, Device, and Equipment for an Automatic Prescription Moxibustion Robot." This invention establishes a complete acupoint framework diagram based on a client's characteristic points, compares this diagram with existing database templates, automatically selects and matches the most suitable acupoint template, and then determines the moxibustion plan based on the client's needs. Through machine selection and filtering, it matches the most suitable plan and executes it automatically, achieving the effect of machine prescription without physician intervention, thus realizing self-service moxibustion. Therefore, in existing technologies, self-service moxibustion robots for the general public can, under guided control, replace doctors in performing certain moxibustion actions, making the moxibustion process more precise and stable.

[0004] As disclosed in Chinese Invention Patent CN114373190B, entitled "An Intelligent Recognition and Automatic Positioning System for Human Acupoints," this invention is based on the anatomical landmark positioning method, bone measurement positioning method, and finger measurement positioning method of Traditional Chinese Medicine acupoint theory. Combined with fractal geometry, it constructs a digital three-dimensional acupoint set, namely a fractal digital meridian map, which can be scaled with high precision to accommodate users of different body types. A 3D camera is used to collect the skin features of the current human body in real time, thereby reconstructing the three-dimensional structural information of the image. Through an information processing system, deep learning technology is used to automatically abstract layer by layer, improving the accuracy of extracting important features of human acupoints. The three-dimensional acupoint set, the current user's three-dimensional posture, and important features are matched and combined to identify and track different acupoints in real time, achieving intelligent acupoint recognition and automatic acupoint location.

[0005] While the aforementioned scheme can be scaled up with high precision to accommodate real-time acupoint identification and tracking for users of different body types, thus achieving intelligent acupoint recognition and automatic acupoint location, in practical applications, the diverse user groups of intelligent moxibustion robots lead to varying representations when collecting skin features. Furthermore, each user's posture is not uniform or fixed, and this cannot be guaranteed to remain constant during subsequent acupoint identification and location. Therefore, the different postures of different users affect feature extraction during the identification process, resulting in a deviation between the identified target location and the actual location, leading to significant errors in the moxibustion therapy positioning. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a calculation method for moxibustion therapy positions, which solves the technical problem of deviations in actual positions caused by changes in posture during moxibustion.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a calculation method for moxibustion therapy locations, the method comprising the following steps: S1. Calibrate the data acquisition device at the end of the moxibustion robot, and collect the user's body point cloud data in real time along the direction of movement, starting from the marked point. ; S2. Retrieve the standard fractal digital meridian diagram and analyze it in the shape point cloud data. Establish a corresponding matching coordinate system on the fractal digital meridian diagram. ; S3, in the matching coordinate system The following is a list of shape point cloud data. Preliminary registration with the fractal digital meridian diagram yields the conversion equation between the two; S4. Determine the shape point cloud data based on the transformation equation. any feature point Relative to the standard acupoints in the fractal digital meridian chart The correction scalar, which includes the rotation matrix. and translation vector ; S5. Apply the correction scalar to the feature points. Positional correction is performed to obtain the target acupoints. The actual location; S6, target acupoints The actual position is fed back to the moxibustion robot system to control the end of the device that carries the moxibustion device to move to the corresponding moxibustion therapy position.

[0008] Furthermore, in step S1, the specific process includes the following steps: S11. Construct a calibration platform, including a binocular vision camera, the end effector of the moxibustion robot, data acquisition equipment, and a moxibustion bed with marker points located in the same space. The binocular camera coordinate system is set as follows: The end-effector coordinate system of the moxibustion robot is The terminal coordinate system of the data acquisition device is ; S12. Set up markers. In the binocular camera coordinate system The coordinates below are Determine the marker point From the terminal coordinate system To the end coordinate system The first transformation relation is: ; And, markers From the end coordinate system To the binocular camera coordinate system The second transformation relation is as follows: ; In the formula, , Terminal coordinate system To the end coordinate system The rotation and translation matrices are as follows; , The end coordinate system is respectively To the binocular camera coordinate system The rotation and translation matrices are as follows; S13. Determine the terminal coordinate system according to the first and second transformation relationships respectively. To the end coordinate system First homogeneous transformation matrix and the end coordinate system To the binocular camera coordinate system The second homogeneous transformation matrix ,Right now: ; ; S14. Merge the first and second transformation relationships to obtain the identifier. From the terminal coordinate system To the binocular camera coordinate system The third transformation relation is: ; ; In the formula, For the marker point The coordinates; S15. Control the end effector of the moxibustion robot to scan the marker points in different postures. The data includes at least four sets of data on the end effector of the moxibustion robot performing only translational movements without any posture changes, and three sets of data on posture changes, resulting in multiple sets of marker points. In the terminal coordinate system coordinates below , and the terminal coordinate system To the binocular camera coordinate system The second homogeneous transformation matrix ; S16, Combine multiple sets of coordinates Second homogeneous transformation matrix Substitute The unknown system of equations is obtained from this, namely: ; In the formula, This indicates that the end effector of the moxibustion robot is controlled to scan the marked points in different postures. The number of groups, ; S17. Solve the unknown system of equations to obtain the rotation matrix. Translation matrix ; S18, Rotation matrix Translation matrix Substitute the first homogeneous transformation matrix In the middle, the terminal coordinate system is obtained. Relative to the end coordinate system Based on the positional relationship, the data collection device at the end of the moxibustion robot is calibrated; S19. After calibration, the end effector of the moxibustion robot collects the user's body point cloud data in real time along the direction of movement, starting from the marked point. .

[0009] Furthermore, in step S17, the specific process includes the following steps: S171. Subtracting the first equation from the other equations in the system of unknown equations in turn yields the first simplified equation, which is: ; S172. Solve the rotation matrix based on the simplified equation. ,Right now: ; S173, Multiple sets of coordinates collected under attitude changes Second homogeneous transformation matrix Substituting the unknowns into the system of equations, we obtain the system of equations by translation, i.e.: ; In the formula, N represents the total number of data collected by the end effector of the moxibustion robot under posture changes. ; S174. Based on the rotation matrix obtained from the solution... Subtracting the remaining equations from the first equation in turn yields the second simplified equation, which is: ; S175. Solve for the translation matrix according to the second simplified equation. ,Right now: ; In the formula, Terminal coordinate system To the end coordinate system The translation matrix below.

[0010] Furthermore, in step S2, the specific process includes the following steps: S21. For the shape point cloud data Preprocessing is performed on the shape point cloud data. Divided into several sub-regions based on human body characteristics. ; S22, in the sub-region The points corresponding to the identified acupoints are selected as feature points. ; S23, Calculate sub-region The corresponding feature points in the current pose Normal vector at point ; S24, Based on normal vector Establish sub-regions Establish a local coordinate system and save the origin coordinates and X, Y, and Z axis vectors of the local coordinate system; S25. Establish the global coordinate system of the fractal digital meridian diagram using the same method as in step S24. ; S26. Select a coordinate system from several local coordinate systems that is compatible with the global coordinate system. The coordinate system with the highest matching degree is used for matching. .

[0011] Furthermore, in step S23, the specific process includes the following steps: S231, Select Individual and feature points Adjacent points that are on the same plane ; S232, Establish Neighboring points Covariance matrix on the plane ,Right now: ; In the above formula, Representing feature points of The three-dimensional centroid of a set of neighboring points; superscript Indicates transpose; S233. Using the PCA algorithm to analyze the covariance matrix. Perform eigenvalue decomposition to obtain the feature points Normal vector The corresponding minimum eigenvalue; S234. Set the viewpoint as the marker point at the starting point of the data acquisition direction. The constraints are as follows: ; In the formula, For feature points Normal vector; These are the viewpoint coordinates; S235, Sub-region The normal vector of all points is set to the line of sight. Direction, used to determine the normal vector at the current pose.

[0012] Furthermore, in step S24, the specific process includes the following steps: S241, in the sub-region China-Israel feature points For fixed point and along the normal vector Direction towards a fixed point Two known acupoints that are closest to each other on either side are randomly selected as points. , ; S242, with fixed points and random points , Construct triangles that lie in the same plane and calculate the lengths of their three sides. , , ; S243, using the normal vector The direction is taken as the Z-axis of the local coordinate system, and according to the edge , , Determine the origin of the local coordinate system as well as axis; like Then use the edge and intersection As the origin of the local coordinate system ,side corresponding vector The direction is Axial direction; like Then use random points or As the origin of the local coordinate system ,side corresponding vector or vector The direction is Axial direction; like Then, with fixed points As the origin of the local coordinate system ,side or edge corresponding vector or vector The direction is Axial direction.

[0013] Further, in step S26, the matching degree The calculation formula is: ; In the above formula, sub-region Middle feature points The coordinates; For the corresponding feature points in the fractal digital meridian diagram Standard acupoints The coordinates; This is a weighted term.

[0014] Further, in step S3, the transformation equation is: ; ; ; In the above formula, For feature points To standard acupoints rotation matrix; For feature points To standard acupoints The translation vector; To match coordinate systems Origin The unit direction vectors of the X, Y, and Z axes; global coordinate system Origin , the unit direction vectors of the X, Y, and Z axes.

[0015] Further, in step S4, the rotation matrix The calculation formula is: ; Translation vector The calculation formula is: ; In the above formula, , Matching coordinate system and global coordinate system The coordinates of the origin.

[0016] By employing the above technical solution, the present invention provides a calculation method for moxibustion therapy locations, which has at least the following beneficial effects: 1. This invention feeds back the corrected actual position coordinates to the moxibustion robot system, enabling the moxibustion robot to quickly locate the corresponding acupoints even when the user's posture changes. Furthermore, during the moxibustion process, when performing techniques such as fixing, pecking, rotating, and moving, the robot can locate the corresponding acupoints without deviation, and autonomously adapt to changes in posture to eliminate corresponding deviations.

[0017] 2. This invention addresses the calibration error of the acquisition device by optimizing the transformation matrix through an improved calibration method, thereby ensuring that the calibration accuracy meets the usage requirements. When the user changes pose, the data acquisition position after the pose change is redefined based on the solved second homogeneous transformation matrix, making it suitable for a diverse range of users. It eliminates the calibration error existing in the operation of the moxibustion robot in the initial stage, and can automatically complete the calibration based on the first homogeneous transformation matrix when facing users in different poses.

[0018] 3. This invention determines the origin O and X-axis by using feature points as fixed points and the direction of the normal vector as the Z-axis of the local coordinate system. Then, it constructs triangles using known acupoint locations. Combined with the local coordinate system with the highest matching degree, it can establish a matching coordinate system with the same direction as the global coordinate system by utilizing the distribution of acupoints. This allows the shape point cloud data and the fractal digital meridian map to be registered in the same coordinate system when they are not registered, thus improving the accuracy of the initial registration between the two. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of the calibration and acquisition device in this invention; Figure 2 This is a schematic diagram of the triangle formation in this invention; Figure 3 In this invention, points are used as the basis for... Origin A schematic diagram of establishing a local coordinate system; Figure 4 In this invention, random points are used. Origin A schematic diagram of establishing a local coordinate system; Figure 5 In this invention, random points are used. Origin A schematic diagram of establishing a local coordinate system; Figure 6 In this invention, a fixed point is used. Origin One of the schematic diagrams for establishing a local coordinate system; Figure 7 The present invention uses a fixed point Origin Schematic diagram of establishing a local coordinate system (Part 2). Detailed Implementation

[0020] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding of how the present application uses technical means to solve technical problems and achieve technical effects, and to facilitate its implementation.

[0021] Self-service moxibustion robots utilize the heat generated by burning mugwort to accurately transfer heat to acupoints on the body through a specific robotic arm and control system. This achieves the traditional Chinese medicine health benefits of unblocking meridians, harmonizing qi and blood, and strengthening the body's resistance to pathogens. However, in practical applications, the diverse user groups of intelligent moxibustion robots lead to variations in the skin characteristics collected. Furthermore, each user's posture is not uniform and fixed, and this cannot be guaranteed to remain constant during subsequent acupoint identification and positioning. Therefore, the different postures of different users affect feature extraction during the identification process, resulting in a deviation between the identified target location and the actual location, leading to significant errors in the moxibustion treatment placement.

[0022] To address the technical issue of significant errors in moxibustion therapy placement due to deviations in actual position caused by changes in posture during the moxibustion process, please refer to... Figures 1-7This embodiment proposes a calculation method for moxibustion therapy positions, enabling the moxibustion robot to quickly locate the corresponding acupoints even when the user's posture changes. Furthermore, during moxibustion, when performing techniques such as fixing, pecking, rotating, and moving, the robot can locate the corresponding acupoints without deviation, and autonomously adapt to changes in posture to eliminate corresponding deviations. The method includes the following steps: S1. Calibrate the data acquisition device at the end of the moxibustion robot, and collect the user's body point cloud data in real time along the direction of movement, starting from the marked point. In the practical application of moxibustion robots, the data collected by the acquisition device is the primary basis for determining the desired moxibustion location. Therefore, the accuracy of the data acquisition device has a significant impact on the final calculation of the moxibustion treatment location. Since the acquisition device is mechanically fixed to the actuator at the end of the moxibustion robot, the main sources of error include errors in the internal structural parameters of the acquisition device and calibration errors. In this embodiment, the acquisition device selected is a line laser sensor or a 3D point cloud scanner, which can directly acquire target point cloud data. The acquisition device used here has already undergone optical plane calibration and parameter calibration, so the errors in the internal structural parameters can be ignored and no further processing is required.

[0023] Calibration error refers to the error between the actual position of the terminal coordinate system and the position obtained through calibration. Calibration error directly causes the acquisition device to deviate from its actual position during data acquisition, and the planned scanning trajectory cannot coincide with the desired position. On the other hand, calibration error affects the accuracy of coordinate transformation after data acquisition. If there is a large calibration error, a significant offset will occur during the transformation from the terminal coordinate system to the end coordinate system after feature point extraction, resulting in a deviation in the final data acquisition route and an inability to match the user's pose. Step S1 specifically includes the following steps: S11. Construct a calibration platform, including a binocular vision camera, the end effector of the moxibustion robot, data acquisition equipment, and a moxibustion bed with marker points located in the same space. The binocular camera coordinate system is set as follows: The end-effector coordinate system of the moxibustion robot is The terminal coordinate system of the data acquisition device is In this embodiment, a binocular vision camera is used as an auxiliary device for the calibration process. Simultaneously, a marker point is pre-set on the moxibustion bed. This marker point can be placed at a location on the moxibustion bed that can be simultaneously captured by both the binocular vision camera and the acquisition device. In this embodiment, the center point of the head placement area on the moxibustion bed is used as the marker point. Figure 1 As shown.

[0024] Here, the marker point coordinates The marker is known. In the binocular camera coordinate system The coordinates are obtained using existing camera calibration methods, including a series of processing steps such as binocular camera calibration, image correction, pixel matching, and 3D coordinate calculation. Specifically: First, based on the matched pixel coordinates and camera parameters, the 3D coordinates of the point in the left and right camera coordinate systems are calculated; then, these two coordinate systems are transformed to the same coordinate system (usually the left camera coordinate system or the world coordinate system) using rotation and translation matrices; finally, the 3D coordinates of the point in the binocular camera coordinate system are obtained. The coordinates are then transformed from the terminal coordinate system... Transform to end coordinate system Then transform to the binocular camera coordinate system. Marker In the binocular camera coordinate system The coordinates below are known, terminal coordinate system To the end coordinate system The transformation relationship is unknown, but it can be obtained by establishing a system of equations.

[0025] S12. Set up markers. In the binocular camera coordinate system The coordinates below are Determine the marker point From the terminal coordinate system To the end coordinate system The first transformation relation is: ; And, markers From the end coordinate system To the binocular camera coordinate system The second transformation relation is as follows: ; In the formula, , Terminal coordinate system To the end coordinate system The rotation and translation matrices are as follows; , The end coordinate system is respectively To the binocular camera coordinate system The rotation and translation matrices are as follows; S13. Determine the terminal coordinate system according to the first and second transformation relationships respectively. To the end coordinate system First homogeneous transformation matrix and the end coordinate system To the binocular camera coordinate system The second homogeneous transformation matrix ,Right now: ; ; S14. Merge the first and second transformation relationships to obtain the identifier. From the terminal coordinate system To the binocular camera coordinate system The third transformation relation is: ; ; In the formula, For the marker point The coordinates; Due to the second homogeneous transformation matrix Marker points coordinates Marker points In the binocular camera coordinate system coordinates below All are known, and the first homogeneous transformation matrix Since the unknown quantity is a system of equations, it is necessary to solve for it. The specific process is as follows: S15. Control the end effector of the moxibustion robot to scan the marker points in different postures. The data includes at least four sets of data on the end effector of the moxibustion robot performing only translational movements without any posture changes, and three sets of data on posture changes, resulting in multiple sets of marker points. In the terminal coordinate system coordinates below , and the terminal coordinate system To the binocular camera coordinate system The second homogeneous transformation matrix Since only translation is performed during data acquisition, the pose of the acquisition device remains unchanged, only its position moves. Therefore, the rotation matrix... Keep it unchanged, and transfer the data. Group data .

[0026] S16, Combine multiple sets of coordinates Second homogeneous transformation matrix Substitute The unknown system of equations is obtained from this, namely: ; In the formula, This indicates that the end effector of the moxibustion robot is controlled to scan the marked points in different postures. The number of groups, ; S17. Solve the unknown system of equations to obtain the rotation matrix. Translation matrix In step S17, the specific process includes the following steps: S171. Subtract the first equation from the other equations in the unknown equation set to obtain the first simplified equation. Since the pose of the acquisition device remains unchanged and only its position moves, the forward motion, which only involves translation and whose posture does not change, is the first simplified equation. Rotation matrix in the system of equations Remain unchanged, while , and Since all quantities are known, the first simplified equation is: ; S172. Solve the rotation matrix based on the simplified equation. ,Right now: ; S173, Multiple sets of coordinates collected under attitude changes Second homogeneous transformation matrix Substituting the unknowns into the system of equations, we obtain the system of equations by translation, i.e.: ; In the formula, N represents the total number of data collected by the end effector of the moxibustion robot under posture changes. ; S174. Based on the rotation matrix obtained from the solution... Subtracting the remaining equations from the first equation in turn yields the second simplified equation. Since... , , and All data are known quantities, meaning they include at least three sets of attitude change data. and Therefore, the second simplified equation is: ; S175. Solve for the translation matrix according to the second simplified equation. ,Right now: ; In the formula, Terminal coordinate system To the end coordinate system The translation matrix below.

[0027] In this embodiment, multiple sets of data are collected by controlling the end effector of the moxibustion robot to perform only translational movements without any posture changes, as well as movements with posture changes. The rotation matrix during translational movements is then utilized. Keeping the equations unchanged, while using known quantities to simplify the unknown equations and solve for the terminal coordinate system. To the end coordinate system Rotation matrix below Then use the rotation matrix obtained from the solution. Use known quantities to simplify the unknown system of equations and solve for the translation matrix. This solution method simplifies the unknown equations to the greatest extent, improves calculation speed, and ensures data diversity by handling multiple translational and attitude change motions, while also improving the handling of rotation matrices. Translation matrix The accuracy of the solution.

[0028] S18, Rotation matrix Translation matrix Substitute the first homogeneous transformation matrix In the middle, the terminal coordinate system is obtained. Relative to the end coordinate system Based on the positional relationship, the data collection device at the end of the moxibustion robot is calibrated; S19. After calibration, the end effector of the moxibustion robot collects the user's body point cloud data in real time along the direction of movement, starting from the marked point. .

[0029] It should be further explained that, in this embodiment, the calibration of the acquisition device is performed by calculating the rotation matrix in conjunction with the binocular vision camera for the first time. Translation matrix These two unknowns determine the first homogeneous transformation matrix. After that, it is no longer necessary to perform the same operation steps every time; it is only necessary to use the determined and solved first homogeneous transformation matrix. The calibration process can be completed. This is mainly done during the first use of the moxibustion robot or after a certain period of operation. Of course, in order to maintain a high standard of accuracy, calibration can also be performed every time it is used. Therefore, the calibration program can be adjusted according to the actual application needs and is not limited to one specific program.

[0030] This embodiment addresses the calibration error of the acquisition device by employing an improved calibration method and optimizing the transformation matrix to achieve the required calibration accuracy. It can redefine the data acquisition position after the user changes pose based on the solved second homogeneous transformation matrix, thus making it suitable for diverse user groups. This eliminates calibration errors present in the initial operation of the moxibustion robot and, when facing users in different poses, can utilize the first homogeneous transformation matrix... Calibration is completed automatically.

[0031] This embodiment calibrates the data acquisition device at the end of the moxibustion robot, eliminating the error between the actual position of the acquisition device and the position obtained through calibration. This prevents trajectory deviation during data acquisition, which could cause the pre-set acquisition trajectory to fail to coincide with the expected trajectory, severely affecting the acquired shape point cloud data. To improve the accuracy, this embodiment uses the above-mentioned calibration method to achieve the transformation between the end coordinate system and the terminal coordinate system. The transformation relationship between the two coordinate systems is determined by the first homogeneous transformation matrix. When the user changes pose, the data acquisition position after the current pose change can be redefined according to the solved first homogeneous transformation matrix, so that it can be applied to users with diverse groups and improve the accuracy of data acquisition.

[0032] S2. Retrieve the standard fractal digital meridian diagram and analyze it in the shape point cloud data. Establish a corresponding matching coordinate system on the fractal digital meridian diagram. Fractal digital meridian mapping is an attempt to combine fractal theory with traditional Chinese medicine meridian theory. It seeks to describe and display the meridian system in a digital form using fractal methods. This atlas not only shows the direction of the meridians and the location of acupoints, but also uses fractal principles to reveal the complexity and self-similarity within the meridian system. As disclosed in Chinese invention patent CN114373190B, this invention, based on the anatomical landmark positioning method, bone measurement positioning method, and finger measurement positioning method of traditional Chinese medicine acupoint theory, combined with fractal geometry, constructs a digital three-dimensional set of human acupoints, namely, a fractal digital meridian map. In step S2, the specific process includes the following steps: S21. For the shape point cloud data Preprocessing is performed on the shape point cloud data. Divided into several sub-regions based on human body characteristics. , representing shape point cloud data The first after the division Sub-regions Preprocessing includes conventional point cloud data processing such as point cloud filtering, downsampling, and outlier removal. Preprocessing retains valid points while removing a large number of invalid points, reducing the computational load of point cloud processing, improving computational efficiency, shortening the processing cycle, and ensuring the real-time performance of the algorithm.

[0033] In this embodiment, the shape point cloud data Divided into several sub-regions based on human body characteristics. Its main purpose is to collect point cloud data of the entire shape that reflects the user's contour in the current pose. Based on fixed human body characteristics, it is divided into multiple sub-regions. A subregion Corresponding to a portion of the human body contour, such as the limb key point detection module, philtrum spine location extraction module, human bone measurement reference extraction module, Dazhui acupoint location extraction module, left and right Jianjing acupoint location extraction module, Shenshu acupoint location extraction module, and Dachangshu acupoint location extraction module disclosed in the prior art, in this embodiment, for example, the areas where the limb key points, philtrum spine, human bone measurement reference, Dazhui acupoint, left and right Jianjing acupoints, Shenshu acupoint, and Dachangshu acupoint are located are respectively the corresponding sub-regions. However, for sub-regions The division can also be changed according to actual application needs, with the aim of dividing the whole into multiple sub-regions with significant characteristics, which will not be elaborated on here.

[0034] S22, in the sub-region The points corresponding to the identified acupoints are selected as feature points. Specifically, according to the above-mentioned division criteria, each sub-region There is at least one designated acupoint, such as Dazhui (GV14), Jianjing (GB21) on both sides, Shenshu (BL23) and Dachangshu (BL25), etc. Therefore, the location of the designated acupoint is known in the fractal digital meridian map. Although changes in the user's posture may cause deviations from this known location, the user is still on the moxibustion bed and in a posture that meets the relative requirements. Therefore, during the establishment of the matching coordinate system, the designated acupoint is still located in the sub-region. Within this framework, after obtaining the correction scalar, the feature points can then be processed. The position is corrected to determine the actual position after the pose change, and the corresponding point can then be used as a feature point.

[0035] S23, Calculate sub-region The corresponding feature points in the current pose Normal vector at point Since the human body's contours (including both clothed and unclothed forms) are represented as curved surfaces, the normal vector is crucial in the subsequent establishment of the matching coordinate system. The tangent plane of the surface at that point is defined, and its direction is perpendicular to the tangent plane of the surface at that point. It is an important reflection of the geometric characteristics of the surface, therefore, it is necessary to define the sub-region. Normal vector at viewpoint The viewpoint is the marker point of the starting point of the aforementioned data acquisition direction. Step S23 specifically includes the following steps: S231, Select Individual and feature points Adjacent points that are on the same plane ;because All neighboring points lie in the same plane, so the variation of neighboring points in the normal direction of this plane is minimal.

[0036] S232, Establish Neighboring points Covariance matrix on the plane ,Right now: ; In the above formula, Representing feature points of The three-dimensional centroid of a set of neighboring points; superscript Indicates transpose; S233. Using the PCA algorithm to analyze the covariance matrix. Perform eigenvalue decomposition to obtain the feature points Normal vector The corresponding minimum eigenvalue; S234. Set the viewpoint as the marker point at the starting point of the data acquisition direction. The constraints are as follows: ; In the formula, For feature points Normal vector; These are the viewpoint coordinates; S235, Sub-region The normal vector of all points is set to the line of sight. Direction, used to determine the normal vector at the current pose.

[0037] In this embodiment, by calculating the normal vectors of feature points, the surface features of the human body at both planar and curved surfaces can be characterized. This allows for the determination of surface features and orientations within a matching coordinate system, and the normal vectors are required to define these features. The Z-axis serves as the matching coordinate system, thus fulfilling the requirements for subsequent coordinate system establishment.

[0038] S24, Based on normal vector Establish sub-regions Establish a local coordinate system and save the origin coordinates and X, Y, and Z axis vectors of the local coordinate system; in step S24, the specific process includes the following steps: S241, in the sub-region China-Israel feature points For fixed point and along the normal vector Direction towards a fixed point Two known acupoints that are closest to each other on either side are randomly selected as points. , ; S242, with fixed points and random points , Construct triangles that lie in the same plane and calculate the lengths of their three sides. , , ; S243, using the normal vector The direction is taken as the Z-axis of the local coordinate system, and according to the edge , , Determine the origin of the local coordinate system as well as axis; like Then use the edge and intersection As the origin of the local coordinate system ,side corresponding vector The direction is Axial direction, such as Figure 3 As shown.

[0039] like Then use random points or As the origin of the local coordinate system ,side corresponding vector or vector The direction is Axial direction, such as Figure 4 , Figure 5 As shown.

[0040] like Then, with fixed points As the origin of the local coordinate system ,side or edge corresponding vector or vector The direction is Axial direction, such as Figure 6 , Figure 7 As shown.

[0041] In this embodiment, by using feature points For fixed point and with normal vector The direction is used as the Z-axis of the local coordinate system. Then, the origin O and X-axis are determined by constructing triangles using the known acupoint locations. Combined with the local coordinate system with the highest matching degree, this allows the distribution of acupoints to be used to establish a coordinate system with the global coordinate system. Matching coordinate systems with the same orientation Thus, the shape point cloud data When the fractal digital meridian diagram is not registered, it can be completed in the same coordinate system, which improves the accuracy of the initial registration between the two.

[0042] S25. Establish the global coordinate system of the fractal digital meridian diagram using the same method as in step S24. Since the fractal digital meridian diagram database is pre-established and stored, the spatial parameters of the corresponding distribution characteristics on it are known in the same coordinate system. In order to reduce the calculation time, the global coordinate system on the fractal digital meridian diagram can be calculated offline and saved.

[0043] S26. Select a coordinate system from several local coordinate systems that is compatible with the global coordinate system. The coordinate system with the highest matching degree is used for matching. In this embodiment, a local coordinate system and a global coordinate system are used. The Euclidean distance between corresponding points is used as a measure of the matching degree between two coordinate systems. The smaller the Euclidean distance, the greater the matching degree. However, directly using the Euclidean distance as the matching degree can easily lead to mismatches between two points, i.e., it is impossible to distinguish them from feature points. For adjacent similar points, this embodiment uses the angle between the intersection of the origin O of the local coordinate system and the side containing the triangle and the Y-axis. Take the integer part and weight it, such as Figure 3 As shown, the included angle obtained through calculation The cosine value is relatively accurate, and after being rounded to the nearest integer, it is used as a weighting term, thus eliminating the aforementioned shortcomings and achieving a precise measurement of the degree of matching between the two coordinate systems. (Degree of matching) The calculation formula is: ; In the above formula, sub-region Middle feature points The coordinates; For the corresponding feature points in the fractal digital meridian diagram Standard acupoints The coordinates; This is a weighted term.

[0044] S3, in the matching coordinate system The following is a list of shape point cloud data. Preliminary registration with the fractal digital meridian diagram yields the transformation equation between the two, namely: ; ; ; In the above formula, For feature points To standard acupoints rotation matrix; For feature points To standard acupoints The translation vector; To match coordinate systems Origin The unit direction vectors of the X, Y, and Z axes; global coordinate system Origin , the unit direction vectors of the X, Y, and Z axes.

[0045] S4. Determine the shape point cloud data based on the transformation equation. any feature point Relative to the standard acupoints in the fractal digital meridian chart The correction scalar, which includes the rotation matrix. and translation vector ; Rotation matrix The calculation formula is: ; Translation vector The calculation formula is: ; In the above formula, , Matching coordinate system and global coordinate system The coordinates of the origin; S5. Apply the correction scalar to the feature points. Positional correction is performed to obtain the target acupoints. The actual position; in this embodiment, when the user's pose does not change, the rotation matrix Translation vector If all values ​​are 0, then the corresponding acupoints correspond one-to-one with the fractal digital meridian diagram, and no correction or compensation is needed. However, when the user's pose changes, the rotation matrix... Translation vector It is no longer 0, at this point it is necessary to... The location is corrected to obtain the target acupoint. ,Right now .

[0046] S6, target acupoints The actual position is fed back to the moxibustion robot system to control the movement of the end effector carrying the moxibustion device to the corresponding moxibustion treatment position. By feeding back the corrected actual position coordinates to the moxibustion robot system, the moxibustion robot can still quickly locate the corresponding acupoints when the user's posture changes. Furthermore, during the moxibustion process, when performing techniques such as fixing, pecking, rotating, and moving, it can locate the corresponding acupoints without deviation, and autonomously adapt to changes in posture to eliminate corresponding deviations.

[0047] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0048] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Since the above embodiments are substantially similar to the method embodiments, their descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0049] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A calculation method for moxibustion therapy locations, characterized in that, The method includes the following steps: S1. Calibrate the data acquisition device at the end of the moxibustion robot, and collect the user's body point cloud data in real time along the direction of movement, starting from the marked point. ; S2. Retrieve the standard fractal digital meridian diagram and analyze it in the shape point cloud data. Establish a corresponding matching coordinate system on the fractal digital meridian diagram. ; S3, in the matching coordinate system The following is a list of shape point cloud data. Preliminary registration with the fractal digital meridian diagram yields the conversion equation between the two; S4. Determine the shape point cloud data based on the transformation equation. any feature point Relative to the standard acupoints in the fractal digital meridian chart The correction scalar, which includes the rotation matrix. and translation vector ; S5. Apply the correction scalar to the feature points. Positional correction is performed to obtain the target acupoints. The actual location; S6, target acupoints The actual position is fed back to the moxibustion robot system to control the end of the device that carries the moxibustion device to move to the corresponding moxibustion therapy position.

2. The calculation method according to claim 1, characterized in that, In step S1, the specific process includes the following steps: S11. Construct a calibration platform, including a binocular vision camera, the end effector of the moxibustion robot, data acquisition equipment, and a moxibustion bed with marker points located in the same space. The binocular camera coordinate system is set as follows: The end-effector coordinate system of the moxibustion robot is The terminal coordinate system of the data acquisition device is ; S12. Set up markers. In the binocular camera coordinate system The coordinates below are Determine the marker point From the terminal coordinate system To the end coordinate system The first transformation relation is: ； And, markers From the end coordinate system To the binocular camera coordinate system The second transformation relation is as follows: ； In the formula, , Terminal coordinate system To the end coordinate system The rotation and translation matrices are as follows; , The end coordinate system To the binocular camera coordinate system The rotation and translation matrices are as follows; S13. Determine the terminal coordinate system according to the first and second transformation relationships respectively. To the end coordinate system First homogeneous transformation matrix and the end coordinate system To the binocular camera coordinate system The second homogeneous transformation matrix ,Right now: ； ； S14. Merge the first and second transformation relationships to obtain the identifier. From the terminal coordinate system To the binocular camera coordinate system The third transformation relation is: ； ； In the formula, For the marker point The coordinates; S15. Control the end effector of the moxibustion robot to scan the marker points in different postures. The data includes at least four sets of data on the end effector of the moxibustion robot performing only translational movements without any posture changes, and three sets of data on posture changes, resulting in multiple sets of marker points. In the terminal coordinate system coordinates below , and the terminal coordinate system To the binocular camera coordinate system The second homogeneous transformation matrix ; S16, Combine multiple sets of coordinates Second homogeneous transformation matrix Substitute The unknown system of equations is obtained from this, namely: ； In the formula, This indicates that the end effector of the moxibustion robot is controlled to scan the marked points in different postures. The number of groups, ; S17. Solve the unknown system of equations to obtain the rotation matrix. Translation matrix ; S18, Rotation matrix Translation matrix Substitute the first homogeneous transformation matrix In the middle, the terminal coordinate system is obtained. Relative to the end coordinate system Based on the positional relationship, the data collection device at the end of the moxibustion robot is calibrated; S19. After calibration, the end effector of the moxibustion robot collects the user's body point cloud data in real time along the direction of movement, starting from the marked point. .

3. The calculation method according to claim 2, characterized in that, In step S17, the specific process includes the following steps: S171. Subtracting the first equation from the other equations in the system of unknown equations in turn yields the first simplified equation, which is: ； S172. Solve the rotation matrix based on the simplified equation. ,Right now: ； S173, Multiple sets of coordinates collected under attitude changes Second homogeneous transformation matrix Substituting the unknowns into the system of equations, we obtain the system of equations by translation, i.e.: ； In the formula, N represents the total number of data collected by the end effector of the moxibustion robot under posture changes. ; S174. Based on the rotation matrix obtained from the solution... Subtracting the remaining equations from the first equation in turn yields the second simplified equation, which is: ； S175. Solve for the translation matrix according to the second simplified equation. ,Right now: ； In the formula, Terminal coordinate system To the end coordinate system The translation matrix below.

4. The calculation method according to claim 1, characterized in that, In step S2, the specific process includes the following steps: S21. For the shape point cloud data Preprocessing is performed on the shape point cloud data. Divided into several sub-regions based on human body characteristics. ; S22, in the sub-region The points corresponding to the identified acupoints are selected as feature points. ; S23, Calculate sub-region The corresponding feature points in the current pose Normal vector at point ; S24, Based on normal vector Establish sub-regions Establish a local coordinate system and save the origin coordinates and X, Y, and Z axis vectors of the local coordinate system; S25. Establish the global coordinate system of the fractal digital meridian diagram using the same method as in step S24. ; S26. Select a coordinate system from several local coordinate systems that is compatible with the global coordinate system. The coordinate system with the highest matching degree is used for matching. .

5. The calculation method according to claim 4, characterized in that, In step S23, the specific process includes the following steps: S231, Select Individual and feature points Adjacent points that are on the same plane ; S232, Establish Neighboring points Covariance matrix on the plane ,Right now: ； In the above formula, Representing feature points of The three-dimensional centroid of a set of neighboring points; superscript Indicates transpose; S233. Using the PCA algorithm to analyze the covariance matrix. Perform eigenvalue decomposition to obtain the feature points Normal vector The corresponding minimum eigenvalue; S234. Set the viewpoint as the marker point at the starting point of the data acquisition direction. The constraints are as follows: ； In the formula, For feature points Normal vector; These are the viewpoint coordinates; S235, Sub-region The normal vector of all points is set to the line of sight. Direction, used to determine the normal vector at the current pose.

6. The calculation method according to claim 4, characterized in that, In step S24, the specific process includes the following steps: S241, in the sub-region China-Israel feature points For fixed point and along the normal vector Direction towards a fixed point Two known acupoints that are closest to each other on either side are randomly selected as points. , ; S242, with fixed points and random points , Construct triangles that lie in the same plane and calculate the lengths of their three sides. , , ; S243, using the normal vector The direction is taken as the Z-axis of the local coordinate system, and according to the edge , , Determine the origin of the local coordinate system as well as axis; like Then use the edge and intersection As the origin of the local coordinate system ,side corresponding vector The direction is Axial direction; like Then use random points or As the origin of the local coordinate system ,side corresponding vector or vector The direction is Axial direction; like Then, using a fixed point As the origin of the local coordinate system ,side or edge corresponding vector or vector The direction is Axial direction.

7. The calculation method according to claim 4, characterized in that, In step S26, the matching degree The calculation formula is: ； In the above formula, sub-region Middle feature points The coordinates; For the corresponding feature points in the fractal digital meridian diagram Standard acupoints The coordinates; This is a weighted term.

8. The calculation method according to claim 1, characterized in that, In step S3, the transformation equation is: ； ； ； In the above formula, For feature points To standard acupoints rotation matrix; For feature points To standard acupoints The translation vector; To match coordinate systems Origin The unit direction vectors of the X, Y, and Z axes; global coordinate system Origin , the unit direction vectors of the X, Y, and Z axes.

9. The calculation method according to claim 8, characterized in that, In step S4, the rotation matrix The calculation formula is: ； Translation vector The calculation formula is: ； In the above formula, , Matching coordinate system and global coordinate system The coordinates of the origin.

Images