Danger avoidance path planning method and system for physiotherapy equipment

Abstract

The invention discloses a danger avoidance path planning method and system for physiotherapy equipment, and the method comprises the steps: S1, carrying out the image collection, preprocessing, feature extraction and three-dimensional space conversion of human veins, and generating a vein risk map; s2, dynamically capturing a change track of a human body posture, and setting a safety threshold value of human body posture change; s3, generating a motion track of the physiotherapy equipment according to the coordinates of the human body target physiotherapy area, the vein risk map and the kinematics parameters of the physiotherapy equipment; and S4, judging whether the human body posture adjustment range exceeds a safety threshold, and if so, re-planning the motion track of the physiotherapy equipment. According to the method, near-infrared vein development data is converted into a quantifiable three-dimensional risk map through a closed-loop architecture of near-infrared vein development, three-dimensional reconstruction, dynamic planning and feedback execution, a vascular safety threshold dynamic model is embedded in path planning, and a vascular safety threshold is obtained through a hybrid algorithm of global RRT and a local artificial potential field method. The physiotherapy robot can accurately avoid dangerous structures such as subcutaneous veins.

Description

TECHNICAL FIELD

[0002] The present application relates to the technical field of path planning of physiotherapy equipment, and in particular relates to a dangerous avoidance path planning method and system for physiotherapy equipment. BACKGROUND

[0004] In recent years, the combination of medical imaging (such as near-infrared vein imaging) and robot path planning algorithm has become a hot topic. The vein imaging technology enhances the visualization of subcutaneous veins through near-infrared light, helping the robot to identify the blood vessel distribution in real time, while deep learning and traditional algorithms optimize the obstacle avoidance path.

[0005] Chinese patent CN116825282A (application date: June 30, 2023) discloses a physiotherapy method for a physiotherapy robot based on visual positioning, which can calculate the local body pose transformation of the physiotherapy object through feature matching without affecting the work of the physiotherapy mechanical arm, update the spatial position of the acupoint, and timely correct the physiotherapy trajectory, better solving the position loss problem of the acupoint when it is blocked and displaced during the work of the physiotherapy robot. However, this method can only display the spatial position of the acupoint, and can only update the spatial position of the acupoint according to the body pose transformation, lacking the display modeling and avoidance ability of subcutaneous fragile tissues (such as superficial veins), so the present application proposes a path planning method for human structure danger avoidance based on vein imaging. SUMMARY

[0007] In view of the problems existing in the prior art, the present application provides a dangerous avoidance path planning method for physiotherapy equipment.

[0008] To achieve the above-mentioned purpose, the specific scheme of the present application is as follows:

[0009] A dangerous avoidance path planning method for physiotherapy equipment, comprising the following steps: S1: image acquisition, preprocessing, feature extraction and three-dimensional space conversion are performed on human veins to generate a vein risk map; S2: dynamically capture the change trajectory of the human body posture, and set a safety threshold for the change of the human body posture; S3: generate a motion trajectory of the physiotherapy equipment according to the coordinates of the target physiotherapy area of the human body, the vein risk map and the kinematic parameters of the physiotherapy equipment; S4: judge whether the range of human body posture adjustment exceeds the safety threshold, if it exceeds the safety threshold, re-plan the motion trajectory of the physiotherapy equipment.

[0010] Preferably, in the step S1, the image acquisition method is to collect the original data of human vein imaging by using near-infrared light imaging method.

[0011] Preferably, in the step S1, the feature extraction method is to fuse multi-scale convolution and Swin-Transformer model to segment the vein contour and extract the vein diameter, branch point and direction.

[0012] Preferably, in the step S1, the three-dimensional space conversion method is to convert the two-dimensional coordinates of the vein image into three-dimensional coordinates by combining the coordinate information of the depth camera.

[0013] Preferably, in the step S2, the change trajectory of the human body posture is obtained by real-time capture of the depth camera.

[0014] Preferably, in the step S2, the safety threshold is dynamically adjusted according to the physiotherapy scene.

[0015] Preferably, in the step S3, the motion trajectory of the physiotherapy device includes a global path and a local obstacle avoidance path, the global path is generated by using the RRT algorithm, and the local obstacle avoidance path is generated by using the artificial potential field method.

[0016] Preferably, in the step S4, the re-planning of the motion trajectory of the physiotherapy device includes steps S1, S2, S3 and S4.

[0017] The application also discloses a risk avoidance path planning system of a physiotherapy device, which comprises: An image acquisition module is configured to acquire real-time human vein images. An image processing module is configured to pre-process the human vein images. A feature extraction module is configured to extract features of the human vein images. A human posture tracking module is configured to dynamically capture the change trajectory of the human posture. A risk area modeling module is configured to convert the human vein images into three-dimensional space and generate a vein risk map. A safety threshold module is configured to set and adjust the safety threshold of the change of the human posture. A path planning module is configured to generate the motion trajectory of the physiotherapy device. An execution control module is configured to control the physiotherapy device to execute the motion trajectory. A device execution module is configured to execute the motion trajectory and feed back motion data in real time. A feedback correction module is configured to correct the trajectory parameters according to the vein images and the motion data, and feed back the corrected parameters to the path planning module. The image acquisition module, the image processing module, the feature extraction module, the risk area modeling module, the path planning module, the execution control module, the device execution module and the feedback correction module are connected in sequence, the human body posture tracking module is connected with the risk area modeling module, the path planning module is further connected with the safety threshold module and the feedback correction module, and the image acquisition module is further connected with the feedback correction module.

[0018] Preferably, the physiotherapy device comprises a multi-axis mechanical arm, the end of the mechanical arm is provided with a physiotherapy probe and a depth camera 13, and the physiotherapy device is further provided with an IMU (inertial measurement unit).

[0019] The technical scheme of the present application has the following beneficial effects:

[0020] The present application realizes accurate avoidance of dangerous structures such as subcutaneous veins by the physiotherapy robot through the closed-loop architecture of "near-infrared vein imaging-three-dimensional reconstruction-dynamic planning-execution feedback", converts the near-infrared vein imaging data into a quantifiable three-dimensional risk map, embeds a blood vessel safety threshold dynamic model in path planning, and realizes local trajectory re-planning caused by human body micro-movement through a hybrid algorithm of global RRT and local artificial potential field method, thereby solving the defects of the traditional scheme that only avoids body surface obstacles and has poor dynamic adaptability. In actual application, the risk of vein injury can be reduced, and the physiotherapy operation efficiency can be improved, and finally the comprehensive effect of "safe redundancy controllable, accurate and efficient operation, and dynamic self-adaptation to human body changes" in a high-risk near-body physiotherapy scene is realized. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 The flowchart of the present application is shown in the figure; Figure 2 The functional module block diagram of the present application is shown in the figure; Figure 3 The schematic diagram of the physiotherapy device of the present application is shown in the figure. DETAILED DESCRIPTION

[0023] The present application is further described below in combination with the drawings and specific embodiments.

[0024] Referring to Figure 1 The present application provides a dangerous avoidance path planning method for a physiotherapy device, comprising the following steps: S1: image acquisition, preprocessing, feature extraction and three-dimensional space conversion are performed on human body veins to generate a vein risk map; S2: dynamically capture the change trajectory of the human body posture, and set the safety threshold of the human body posture change; S3: generate the motion trajectory of the physiotherapy device according to the coordinates of the human body target physiotherapy area, the vein risk map and the kinematics parameters of the physiotherapy device. S4: judging whether the range of human posture adjustment exceeds a safety threshold, and if so, re-planning the motion trajectory of the physiotherapy device.

[0025] In the step S1, the method of image acquisition is to collect the original data of human vein imaging by using near-infrared light imaging method, which solves the problem that the traditional physical obstacle avoidance cannot perceive subcutaneous blood vessels.

[0026] In the step S1, the method of feature extraction is to segment the vein contour by fusing multi-scale convolution and Swin-Transformer model, and to extract the pipe diameter, branch point and direction of the vein.

[0027] In the step S1, the method of three-dimensional space conversion is to convert the two-dimensional coordinates of the vein image into three-dimensional coordinates by combining the coordinate information of the depth camera, and to determine the depth of the vein blood vessel under the skin.

[0028] In the step S2, the changing trajectory of the human posture is obtained by real-time capture by the depth camera, the human posture is captured in real time by the depth camera, and the motion trajectory of the key part of the human body (such as the physiotherapy area) is output; solve the influence of the patient's dynamic body position on the position of the vein, ensure that the three-dimensional coordinates of the vein are updated in real time with the movement of the human body, and avoid the failure of avoidance caused by static planning.

[0029] In the step S2, the safety threshold is dynamically adjusted according to the physiotherapy scene, the blood vessel features are converted into a "space forbidden zone" that can be understood by the mechanical arm, the "physiological risk quantization" is realized, and clear avoidance rules are provided for path planning.

[0030] In the step S3, the motion trajectory of the physiotherapy device includes a global path and a local obstacle avoidance path, the method of generating the global path adopts RRT algorithm, and the method of generating the local obstacle avoidance path adopts artificial potential field method, so as to generate a collision-free trajectory.

[0031] In the step S4, the re-planning of the motion trajectory of the physiotherapy device includes steps S1, S2, S3 and S4, that is, when the position of the vein deviates due to the movement of the human body and exceeds the safety threshold, the vein image needs to be re-collected, a new motion trajectory is planned, and the motion smoothness is ensured.

[0032] In the process of executing the motion trajectory, the physiotherapy device is driven by the PID control algorithm, and the error is corrected by closed-loop feedback, so as to realize the goal of "precise operation + safe avoidance".

[0033] Referring to Figure 2 , the present application also provides a dangerous avoidance path planning system for a physiotherapy device, comprising: An image acquisition module 1 is configured to acquire real-time images of human veins, and the image acquisition is achieved by using a near-infrared light imaging method to provide original visual data of vein distribution and solve the problem that a traditional physical obstacle avoidance cannot perceive subcutaneous blood vessels. An image processing module 2 is configured to pre-process the images of human veins. A feature extraction module 3 is configured to extract features of the images of human veins, and a deep learning model (a model combining multi-scale convolution and Swin-Transformer) is used to segment vein contours and extract features such as pipe diameter, branch point and direction. A human posture tracking module 4 is configured to dynamically capture the changing trajectory of the human posture, output the motion trajectory of the key parts of the human body (such as the physiotherapy area), solve the influence of the dynamic body position of the patient on the position of the vein, ensure that the three-dimensional coordinates of the vein are updated in real time with the movement of the human body, and avoid the failure of avoidance caused by static planning. A risk area modeling module 5 is configured to combine the coordinate information of the depth camera, convert the images of human veins into three-dimensional space, and then generate a vein risk map. A safety threshold module 6 is configured to set and adjust the safety threshold of the human posture change in combination with the physiotherapy scene, convert the vein vessel features into a "space forbidden zone" that can be understood by the physiotherapy device, realize "physiological risk quantization", and provide clear avoidance rules for path planning. A path planning module 7 is configured to generate the motion trajectory of the physiotherapy device, including a global path and a local obstacle avoidance path. The RRT algorithm is used to generate the global path, and the artificial potential field method is used to generate the local obstacle avoidance path, so as to generate a collision-free trajectory. When the position of the vein deviates due to the movement of the human body and exceeds the safety threshold, the motion trajectory is re-planned to ensure the smoothness of the motion. As the "brain" of the system, the path planning module plans the optimal motion path under the premise of meeting the safety avoidance, and balances safety and operation efficiency. An execution control module 8 is configured to control the physiotherapy device to execute the motion trajectory, convert the planned motion trajectory into motion instructions of the physiotherapy device, and drive the physiotherapy device to execute the motion through a PID control algorithm. A device execution module 9 is configured to execute the motion trajectory and feed back motion data in real time, as the "hands and feet" of the system, and convert the planned instructions into actual motion. A feedback correction module 10 is configured to correct the trajectory parameters according to the vein images and the motion data, feed back the corrected parameters to the path planning module 8, correct errors through a closed-loop feedback, dynamically adjust the trajectory parameters, and finally realize the goal of "precise operation + safety avoidance". The image acquisition module 1, the image processing module 2, the feature extraction module 3, the risk area modeling module 5, the path planning module 7, the execution control module 8, the device execution module 9 and the feedback correction module 10 are connected in turn, the human body posture tracking module 4 is connected with the risk area modeling module 5, the path planning module 8 is also connected with the safety threshold module 6 and the feedback correction module 10, and the image acquisition module 1 is also connected with the feedback correction module 10.

[0034] Referring to Figure 3 The physiotherapy device comprises a multi-axis mechanical arm 11, a physiotherapy probe 12 and a depth camera 13 are installed at the tail end of the mechanical arm 11, and the physiotherapy device is also provided with an IMU (inertial measurement unit), and other mechanical structures of the physiotherapy device can adopt prior art, which will not be described here.

[0035] This embodiment takes the physiotherapy process of a lumbar muscle strain patient as an example to further illustrate the present application.

[0036] Scenario background

[0037] The lumbar muscle strain patient often has muscle tension and pain in the waist and back due to long-term sitting and poor posture, and needs to be treated by the radio frequency massage head of the physiotherapy robot acting on the erector spinae muscle area of the waist and back to relieve muscle spasm. This scenario needs to avoid the subcutaneous veins of the waist and back (such as the branches of the lumbar vein, with a pipe diameter of 2-4 mm and a subcutaneous depth of 3-6 mm) to prevent the radio frequency energy from directly acting on the blood vessels to cause local temperature to be too high to cause hematoma, and to cope with the dynamic body position changes such as patient breathing (trunk fluctuation amplitude of 2-3 mm) and adjusting sitting posture to prevent static path planning from causing the physiotherapy probe (such as the radio frequency massage head) to touch the blood vessels by mistake.

[0038] Working process Vein modeling and safety threshold setting: the near-infrared vein imaging module scans the waist and back, collects vein images, divides the lumbar vein branch contour through the “multi-scale convolution-SwinTransformer model”, quantifies the blood vessel pipe diameter and subcutaneous depth, and determines it as a “medium-risk blood vessel”; Dynamic data acquisition and vein tracking: the depth camera collects the point cloud of the waist and back in real time at a frame rate of 30 Hz, and the IMU (attached to the skin of the waist and back) collects acceleration data, and synchronously captures the trunk fluctuation (Z-axis direction ±2 mm) caused by patient breathing and the waist and back deviation caused by unconscious adjustment of sitting posture; The surface point cloud deformation of the waist and back is transmitted to the vein model through the “non-rigid registration algorithm”, the “surface-vein” deformation transmission coefficient is combined, the three-dimensional coordinates of the vein are corrected in real time (such as the original vein coordinates (x=600, y=300, z=120) are updated to (x=598.5, y=300, z=122)), and the position of the risk vein is adjusted synchronously.

[0039] Dynamic path planning and obstacle avoidance: Global path: using improved RRT algorithm, planning the initial path of the radio frequency massage head from the "standby position (10 cm lateral to the waist and back)" to the "physiotherapy area", the path needs to cover the physiotherapy area and the minimum distance to the forbidden area is greater than or equal to 2mm; Local re-planning: when the vein coordinate update amount is greater than 1mm (such as the patient's deep breathing causes the vein to move up 2mm), trigger the artificial potential field method, recalculate the path within 40ms, take the physiotherapy area grid point as the "gravitational source" and the forbidden area as the "repulsive source", adjust the motion trajectory of the massage head along the combined force direction, and ensure that the position of the risk vein is always avoided.

[0040] Physiotherapy execution and safety monitoring: The mechanical arm drives the radio frequency massage head to move along the planned path, outputs radio frequency energy according to the preset parameters, and monitors the massage head in real time; If the depth camera detects that the distance between the massage head and the vein is less than 4mm (80% of the safety threshold), or the IM detects a sudden body movement (such as the patient suddenly moves sideways), immediately suspend the radio frequency output and motion, and restart after the path is adjusted.

[0041] Core principle:

[0042] Vein visualization and risk conversion: using the high absorption rate of near-infrared light to blood hemoglobin, breaking through the interference of the fat layer of the waist and back to clearly visualize the vein, dividing the risk level by quantifying the vascular parameters, converting "vascular injury risk" into "space forbidden area" that can be recognized by the robot, solving the problem of "unable to avoid obstacles because of invisible blood vessels".

[0043] Dynamic tracking and adaptation: through "surface deformation-vein displacement" mapping and multi-source data fusion (camera+IMU), real-time correlation of human dynamic body position and vein position change, to avoid "vein displacement but path not updated" avoidance failure caused by breathing and body movement; Hierarchical path planning: global path ensures physiotherapy coverage, local re-planning responds to dynamic risks, balances "physiotherapy effect" and "safe obstacle avoidance", and further reduces the probability of blood vessel injury through temperature and distance dual monitoring.

[0044] The above only describes the preferred embodiments of the present application, and does not limit the patent range of the present application, any equivalent structural transformation made according to the inventive concept of the present application, or direct / indirect application in other related technical fields is included in the protection scope of the present application.

Claims

1. A method for planning hazard avoidance paths for physiotherapy equipment, characterized in that, Includes the following steps: S1: Perform image acquisition, preprocessing, feature extraction, and 3D spatial transformation on human veins to generate a vein risk map; S2: Dynamically capture the trajectory of changes in human posture and set a safety threshold for changes in human posture; S3: Generate the motion trajectory of the physiotherapy equipment based on the coordinates of the target physiotherapy area of ​​the human body, the venous risk map, and the kinematic parameters of the physiotherapy equipment; S4: Determine whether the range of human posture adjustment exceeds the safety threshold. If it does, re-plan the motion trajectory of the physiotherapy equipment.

2. The hazard avoidance path planning method for physiotherapy equipment according to claim 1, characterized in that, In step S1, the image acquisition method is to acquire raw data of human vein imaging using near-infrared light imaging.

3. The hazard avoidance path planning method for physiotherapy equipment according to claim 2, characterized in that, In step S1, the feature extraction method is to segment the vein contour using a model that integrates multi-scale convolution and Swin-Transformer, and extract the vein diameter, branch points, and direction.

4. The hazard avoidance path planning method for physiotherapy equipment according to claim 3, characterized in that, In step S1, the method of three-dimensional spatial transformation is to combine the coordinate information of the depth camera to convert the two-dimensional coordinates of the vein image into three-dimensional coordinates.

5. The hazard avoidance path planning method for physiotherapy equipment according to claim 1, characterized in that, In step S2, the trajectory of the human body's posture change is obtained in real time by a depth camera.

6. The hazard avoidance path planning method for physiotherapy equipment according to claim 5, characterized in that, In step S2, the safety threshold is dynamically adjusted according to the physiotherapy scenario.

7. The hazard avoidance path planning method for physiotherapy equipment according to claim 1, characterized in that, In step S3, the motion trajectory of the physiotherapy device includes a global path and a local obstacle avoidance path. The global path is generated using the RRT algorithm, and the local obstacle avoidance path is generated using the artificial potential field method.

8. The hazard avoidance path planning method for physiotherapy equipment according to claim 1, characterized in that, In step S4, replanning the motion trajectory of the physiotherapy equipment includes steps S1, S2, S3, and S4.

9. A hazard avoidance path planning system for physiotherapy equipment, characterized in that, include: The image acquisition module is used to acquire images of human veins in real time. The image processing module is used to preprocess human vein images; The feature extraction module is used to extract features from human vein images; The human posture tracking module is used to dynamically capture the trajectory of changes in human posture; The risk area modeling module is used to transform human vein images into three-dimensional space, thereby generating a vein risk map; The safety threshold module is used to set and adjust the safety threshold for changes in human posture. The path planning module is used to generate the motion trajectory of the physiotherapy equipment; The execution control module is used to control the physiotherapy equipment to execute the motion trajectory; The device execution module is used to execute the motion trajectory and provide real-time feedback of motion data; The feedback correction module corrects trajectory parameters based on vein images and motion data, and feeds the corrected parameters back to the path planning module. The image acquisition module, image processing module, feature extraction module, risk area modeling module, path planning module, execution control module, device execution module, and feedback correction module are connected in sequence. The human posture tracking module is connected to the risk area modeling module. The path planning module is also connected to the safety threshold module and the feedback correction module. The image acquisition module is also connected to the feedback correction module.

10. The hazard avoidance path planning system for physiotherapy equipment according to claim 9, characterized in that, The physiotherapy device includes a multi-axis robotic arm, the end of which is equipped with a physiotherapy probe and a depth camera 13. The physiotherapy device is also equipped with an IMU (inertial measurement unit).

Images