An intelligent rehabilitation device and method based on machine vision and thermotherapy technology
By combining multi-angle machine vision monitoring and intelligent positioning of the thermotherapy head with generative adversarial network assessment, the problem of insufficient motion monitoring and thermotherapy positioning in traditional rehabilitation equipment has been solved, thereby improving the accuracy and comfort of knee joint rehabilitation.
Patent Information
- Application Number
- CN202610480912.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-13
- Publication Date
- 2026-07-10
AI Technical Summary
Traditional rehabilitation equipment lacks objective motion monitoring and precise thermotherapy positioning in the rehabilitation of lower limb joints such as knee osteoarthritis. The synergy between rehabilitation and thermotherapy is insufficient, resulting in difficulty in accurately tracking training effects and inefficient resource utilization.
A multi-angle machine vision monitoring module is used to collect patients' lower limb movement data in real time. Combined with the intelligent positioning and temperature control module of the thermotherapy head, the rehabilitation effect is evaluated by generating an adversarial network, and the real-time collaborative optimization of movement monitoring and thermotherapy parameters is achieved.
It has improved the precision and comfort of knee joint rehabilitation, and enhanced the quantitative assessment of rehabilitation effects and patient compliance through real-time motion monitoring and adjustment of thermotherapy parameters.
Smart Images
Figure CN122350979A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of rehabilitation medicine and medical equipment technology, specifically relating to an intelligent rehabilitation device and method that integrates machine vision and thermotherapy technology. Background Technology
[0002] Traditional rehabilitation equipment has the following problems in the rehabilitation of lower limb joints such as knee osteoarthritis:
[0003] 1. Rehabilitation movement monitoring relies on manual intervention: During rehabilitation training, the assessment of the standardization, continuity, and accuracy of patients' movements largely depends on doctors' subjective observation, lacking objective and real-time quantitative analysis, which makes it difficult to accurately track the training effect. 2. Insufficient positioning and temperature control in thermotherapy: Most existing thermotherapy equipment uses fixed heat sources or manual adjustment, making it difficult to dynamically adjust the position and temperature of the thermotherapy head according to the real-time changes in the patient's limb movements. This results in limited thermotherapy effects, especially in the inability to accurately target key acupoints (such as the inner knee eye and Yanglingquan) to relieve pain. 3. Weak synergy between rehabilitation and thermotherapy: Rehabilitation training and thermotherapy usually operate independently, failing to be coordinated and optimized in conjunction with the patient's movement status and physiological needs (such as pain level), resulting in inefficient resource utilization and poor patient experience.
[0004] Although existing technologies attempt to incorporate machine vision or thermotherapy assistance, their limitations are as follows: 1. Machine vision and hyperthermia functions are not deeply integrated, making it impossible to achieve real-time coordinated adjustment of motion monitoring and hyperthermia parameters; 2. The lack of quantitative assessment and dynamic optimization mechanisms for rehabilitation effects leads to insufficient matching between heat therapy and training.
[0005] Therefore, there is an urgent need for an intelligent rehabilitation device and method that integrates machine vision and thermotherapy technology to solve the above problems through motion recognition, thermotherapy positioning optimization, and dynamic parameter adjustment. Summary of the Invention
[0006] The purpose of this invention is to provide an intelligent rehabilitation device and method based on machine vision and thermotherapy technology, which aims to achieve real-time monitoring of rehabilitation training movements and dynamic optimization of thermotherapy effects through visual recognition and precise control of thermotherapy, thereby improving the accuracy and comfort of lower limb joint (such as knee) rehabilitation.
[0007] To achieve the above objectives, the present invention provides an intelligent rehabilitation device based on machine vision and thermotherapy technology, comprising: A multi-angle machine vision monitoring module is used to collect patients' lower limb movement data in real time and generate dynamic motion feature vectors; The intelligent positioning and temperature control module for the thermotherapy head includes a displacement adjustment mechanism, a thermotherapy head, and a temperature control system. The displacement adjustment mechanism adjusts the spatial position of the thermotherapy head according to the dynamic motion feature vector, and the temperature control system dynamically adjusts the thermotherapy temperature according to the patient's movement state. The assistive flexion and extension mechanism and linkage control module, including the knee assistive brace, strap and drive assembly, are used to assist patients in completing flexion and extension movements; The rehabilitation effect evaluation and parameter optimization module generates standard rehabilitation movement features based on generative adversarial networks, generates a rehabilitation movement index by comparing it with the dynamic movement feature vector, and optimizes the equipment operating parameters accordingly. The real-time feedback and remote monitoring module is used to provide real-time action correction prompts and remote data transmission; The multi-angle machine vision monitoring module, the intelligent positioning and temperature control module of the thermotherapy head, and the auxiliary flexion and extension mechanism and linkage control module achieve data fusion and closed-loop linkage control through the main control chip.
[0008] Preferably, the multi-angle machine vision monitoring module includes three RGB-D cameras arranged in a 120° ring, installed at a height of 60-80cm above the ground, with external parameters aligned using the Zhang Zhengyou calibration method, and a world coordinate system established with the knee joint rotation center as the origin.
[0009] Preferably, the multi-angle machine vision monitoring module uses the HOG algorithm to extract limb contours and joint position features from keyframes, and performs multi-scale spatiotemporal feature fusion based on the Vision Transformer deep learning model to generate the dynamic motion feature vector.
[0010] Preferably, the intelligent positioning and temperature control module of the heat therapy head identifies the three-dimensional coordinates of the knee joint and key acupoints such as the inner knee eye and Yanglingquan based on visual data, and drives the displacement adjustment mechanism to automatically adjust the position of the heat therapy head so that it is aligned with the target acupoint.
[0011] Preferably, the displacement adjustment mechanism consists of a three-axis linear module and a two-degree-of-freedom rotary table. The three-axis linear module has a travel of ±200mm in the X, Y, and Z directions, and the two-degree-of-freedom rotary table can adjust the pitch and yaw by ±30° respectively.
[0012] Preferably, the auxiliary flexion and extension mechanism includes a knee assist bracket, upper and lower leg straps, a knee strap, and a vertical rodless cylinder. The piston rod of the rodless cylinder is mechanically connected to the knee assist bracket, and the main control chip controls the cylinder's extension and retraction speed and stroke based on visual data.
[0013] Preferably, the main control chip integrates the motion trajectory of visual feedback, the temperature status of the thermotherapy head, and the preset rehabilitation plan to adjust the cylinder output force and motion parameters in real time. When the rehabilitation exercise index is lower than the set threshold, it automatically reduces resistance or provides appropriate assistance when the patient's movements are standardized.
[0014] Preferably, the rehabilitation effect assessment and parameter optimization module simulates the standard rehabilitation movement characteristics under ideal conditions through a GAN generation model, generating standard fusion feature data covering normal gait cycle, joint angle change range and limb coordination.
[0015] Preferably, when the rehabilitation exercise index is lower than the set threshold, the main control chip automatically triggers the parameter optimization process, dynamically adjusts the temperature distribution of the heat therapy head and the resistance or assist parameters of the auxiliary flexion and extension mechanism, forming a closed-loop control until the rehabilitation exercise index recovers to the target range.
[0016] The technical solution of the present invention also provides an intelligent rehabilitation method based on machine vision and thermotherapy technology, including the following steps: The multi-angle machine vision monitoring module collects the patient's lower limb movement data in real time, and generates dynamic motion feature vectors by fusing the HOG algorithm with the Transformer model. Based on the dynamic motion feature vector, the coordinates of the knee joint and key acupoints are determined, the displacement adjustment mechanism is controlled to adjust the position of the heat therapy head, and the temperature control system is activated to adjust the temperature of the heat therapy head to the preset range. Based on the generative adversarial network, standard rehabilitation movement features are generated and compared with the dynamic movement feature vector to generate a rehabilitation movement index. The resistance / assistance parameters and thermotherapy temperature of the auxiliary flexion and extension mechanism are dynamically adjusted according to the rehabilitation exercise index to achieve synergistic optimization of rehabilitation training and thermotherapy.
[0017] Compared with the prior art, the present invention has the following beneficial technical effects: 1. Precise motion monitoring and thermotherapy positioning: Real-time capture of the patient's limb movement trajectory and joint position through machine vision, combined with displacement adjustment components to dynamically adjust the position of the thermotherapy head, ensuring that the thermotherapy is precisely applied to the knee joint and key acupoints, thereby improving the pain relief effect; 2. Quantitative assessment of rehabilitation effects: Using GAN to generate standard movement features, a rehabilitation exercise index is generated through similarity calculation to objectively assess the standardization and continuity of rehabilitation movements, reducing the error of doctors' subjective judgment; 3. Dynamic parameter optimization: Based on rehabilitation exercise index and physiological state, the temperature distribution of heat therapy and the resistance / assistance parameters of the auxiliary flexion and extension mechanism are adjusted in real time to achieve synergistic optimization of rehabilitation training and heat therapy, thereby improving treatment efficiency; 4. Improve patient compliance: Real-time feedback and remote monitoring enhance patient engagement; at the same time, the precise positioning of thermotherapy can reduce pain during the rehabilitation process and improve patient compliance. Attached Figure Description
[0018] Figure 1 This is a front view schematic diagram of the overall structure of an intelligent rehabilitation device and method based on machine vision and thermotherapy technology according to the present invention. Figure 2 This is a left-side view of the overall structure of the intelligent rehabilitation device and method based on machine vision and thermotherapy technology according to the present invention. Figure 3 This is a flowchart of the machine vision monitoring process in an intelligent rehabilitation device and method based on machine vision and thermotherapy technology according to the present invention. Figure 4 This is a logic diagram of intelligent positioning and temperature control of the thermotherapy head in an intelligent rehabilitation device and method based on machine vision and thermotherapy technology according to the present invention. Figure 5 This is a flowchart of the rehabilitation effect evaluation and parameter optimization process in an intelligent rehabilitation device and method based on machine vision and thermotherapy technology according to the present invention.
[0019] Attached reference numerals: 1. Camera; 2. Assistive flexion and extension mechanism; 3. Touch screen; 4. Heat therapy head; 5. Displacement adjustment mechanism; 6. Strap; 7. Main control box. Detailed Implementation
[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0021] This invention provides an intelligent rehabilitation device based on machine vision and thermotherapy technology, which includes a multi-angle machine vision monitoring module, a thermotherapy head intelligent positioning and temperature control module, a rehabilitation effect evaluation and parameter optimization module, an auxiliary flexion and extension mechanism and linkage control module, and a real-time feedback and remote monitoring module.
[0022] The overall structure mainly includes components such as the main control box 7, touch screen 3, heat therapy head 4, displacement adjustment mechanism 5, camera 1, assisted flexion and extension mechanism 2, and strap 6. The main control box 7 serves as the electrical and control center of the device, integrating the main control chip, power module, and data processing unit. It is equipped with casters at its bottom for flexible deployment in clinical environments. The touch screen 3, mounted on the top of the device, provides a human-computer interaction interface for displaying rehabilitation data, receiving user commands, and providing real-time visual feedback. Three high-frame-rate RGB-D cameras 1 (preferably Intel RealSense D455) are arranged in a 120° circular pattern on the upper part of the device support, with a height set at 60-80cm from the ground. This height is suitable for the typical knee joint position when the patient is seated. Each camera 1 is aligned using the Zhang Zhengyou calibration method to establish a world coordinate system with the knee joint rotation center as the origin, ensuring that the spatial positioning error is controlled within 1.5mm, thereby achieving high-precision capture of the patient's lower limb movement trajectory.
[0023] In the multi-angle machine vision monitoring module, three high-definition cameras 1 synchronously and in real-time acquire video data of the patient's lower limb rehabilitation training process. This data includes color images and depth information. After receiving the video stream, the control components first perform data acquisition and preprocessing, including noise reduction, frame rate unification, and image enhancement. Then, keyframe localization is performed to extract valid frames containing complete knee joint information from the continuous video stream. In the keyframe processing stage, the HOG (Histogram of Oriented Gradients) algorithm is used to detect the knee joint and surrounding limbs of the target object in the keyframe, locate its motion trajectory and joint position, and extract limb contour features. Specifically, the cell unit size is set to 8×8 pixels, and a gradient histogram in 9 directions is constructed to accurately locate the spatial position of the knee joint and surrounding limbs. The extracted HOG features are then input into a vision-based system. The Transformer (ViT) deep learning model is used for processing. This model adopts a 16×16 pixel patch size and contains a 12-layer encoder structure. It uses a self-attention mechanism to fuse multi-scale spatiotemporal features of video data to generate dynamic motion feature vectors for evaluating the coherence and standardization of rehabilitation movements. In this process, the local texture features of the HOG algorithm and the global temporal features of the Transformer are weighted and fused through an attention weight matrix. This preserves local details and captures the temporal continuity of the movements. The final output dynamic motion feature vector is used to accurately describe the morphological features and movement trajectory of the patient's rehabilitation movements, providing a data foundation for subsequent thermotherapy positioning and rehabilitation assessment.
[0024] In the intelligent positioning and temperature control module of the thermotherapy head, the thermotherapy head 4 achieves mechanical and electrical linkage with the aforementioned machine vision module through a displacement adjustment component, dynamically adjusting the spatial coordinates of the thermotherapy head 4 according to the patient's limb position and joint angle. The displacement adjustment mechanism 5 consists of a three-axis linear module and a two-degree-of-freedom rotary table. The three-axis linear module has a travel of ±200mm in the X, Y, and Z directions, while the two-degree-of-freedom rotary table can achieve pitch and yaw adjustments of ±30° each, with an overall repeatability accuracy of ±0.1mm. When the vision module identifies the three-dimensional coordinates of the patient's knee joint and key acupoints such as the inner knee eye and Yanglingquan, the main control chip drives the displacement adjustment component based on this coordinate information to automatically adjust the spatial position of the thermotherapy head 4, ensuring it accurately aligns with the target acupoints. The thermotherapy head 4 incorporates a semiconductor cooling chip (thermoelectric cooling chip) and a high-precision temperature sensor. The temperature control system employs a PID control algorithm, controlled by a host computer PID controller and... The STM32H7 series main control chip forms a closed-loop feedback control circuit; the temperature control range is 40°C to 60°C, the control accuracy reaches ±0.5°C, the heating rate is controlled to not exceed 5°C per second, and the heat conduction medium uses a medical-grade silicone thermal pad to ensure uniform heat transfer and meet biocompatibility requirements; the data fusion analysis results of the temperature control system and the machine vision module are linked in real time, and the thermotherapy temperature and duration are dynamically adjusted according to the patient's movement status, joint angle, and rehabilitation exercise index. When the patient's movement amplitude is detected to be large or the pain index is high, the temperature is automatically adjusted to a suitable range or the following trajectory of the thermotherapy head 4 is adjusted to ensure that the thermotherapy is always accurately applied to the target area.
[0025] The assisted flexion and extension mechanism and linkage control module include a knee assistive support, a knee cover, upper and lower leg straps, a knee strap, and a linkage structure driven by a vertical rodless cylinder. During implementation, the upper and lower leg straps are fixed to the patient's thigh and calf respectively, and the knee strap secures the patient's knee to the assistive support, forming a stable force transmission path. The vertical rodless cylinder serves as the main driving element, with its piston rod mechanically connected to the knee assistive support. The main control chip controls the cylinder's extension and retraction speed and stroke based on real-time motion data collected by the machine vision module, thereby driving the knee assistive support to move up and down, assisting the patient in completing leg flexion and extension movements. This mechanism achieves three-way linkage with the temperature control system and the machine vision module. The main control chip integrates the visual feedback motion trajectory, the temperature status of the heat therapy head 4, and the preset rehabilitation plan, adjusting the cylinder's output force and motion parameters in real time. When the rehabilitation exercise index is below a set threshold, it automatically reduces resistance to adapt to the patient's current exercise capacity, or provides appropriate assistance when the patient's movements are standardized, ensuring the safety and effectiveness of the training movements.
[0026] The rehabilitation effect assessment and parameter optimization module is based on Generative Adversarial Network (GAN) technology. First, it simulates the characteristics of standard rehabilitation movements under ideal conditions through a GAN generation model, generating standard fusion feature data covering normal gait cycle, joint angle variation range, and limb coordination. Subsequently, the system performs multi-scale fusion and similarity calculation with the dynamic motion feature vector extracted in real time by the machine vision module and the above standard fusion feature data. Specifically, it uses dynamic time warping (DTW) or cosine similarity algorithm to quantify the difference between the two, generating a rehabilitation movement index. This index objectively reflects the standardization and coherence of the patient's movements. The system sets a reasonable range for the rehabilitation movement index. When the detected index is lower than the threshold, the parameter optimization process is automatically triggered. The main control chip dynamically adjusts the temperature distribution of the heat therapy head 4 according to the current index and the patient's physiological state (such as pain level and muscle fatigue), enhances the treatment intensity of key acupoints, and adjusts the resistance or assistance parameters of the auxiliary flexion and extension mechanism 2, such as reducing flexion and extension resistance to adapt to the patient's current ability or adjusting the movement speed to match the patient's tolerance. This optimization process forms a closed-loop control, continuously iterating until the rehabilitation movement index recovers to the target range, thereby achieving synergistic optimization of rehabilitation training and heat therapy effects.
[0027] The real-time feedback and remote monitoring module provides patients with real-time movement correction prompts through the touchscreen smart terminal. When the machine vision module detects that the patient's movements deviate from the standard trajectory, the system prompts the patient to adjust their posture through visual indicators or voice broadcasts. At the same time, all rehabilitation data, including movement trajectory, joint angles, heat therapy temperature, rehabilitation exercise index, and pain assessment results, are uploaded to the cloud server in real time via wireless network. Doctors can access the cloud data through remote terminals to view the patient's rehabilitation progress and detailed reports, and adjust long-term training plans based on historical data trends. The patient and doctor terminals achieve cross-platform collaborative management, supporting doctors to remotely set rehabilitation parameters, and the device to execute medical orders locally and provide feedback on the execution results, forming a continuous management system for in-hospital treatment and home rehabilitation.
[0028] Based on the above-mentioned equipment, this invention also provides an intelligent rehabilitation method based on machine vision and thermotherapy technology, comprising the following steps: 1. Data acquisition and motion recognition: The patient first fixes his lower limb to the assisted flexion and extension mechanism 2 with the strap 6. After the device is started, the multi-angle camera 1 begins to record the lower limb movement process in real time. After the control components receive the visual data, the limb contour and joint position in the key frame are extracted by the HOG algorithm. The Transformer model is then used to perform multi-scale feature fusion to generate a dynamic motion feature vector. 2. Positioning and temperature control of the thermotherapy head: After the vision module identifies the location of the patient's knee joint and surrounding key acupoints, the displacement adjustment component automatically adjusts the position of the heat therapy head 4 according to the heat therapy coordinates determined by the vector. The temperature control system is activated simultaneously to adjust the temperature of the heat therapy head 4 to the preset range. The heat therapy temperature range is 40°C–60°C (accuracy ±0.5°C), and the heating rate is ≤5°C / s. The heat conduction medium is a medical-grade silicone heat-conducting pad. The temperature control system dynamically adjusts the temperature of the heat therapy head 4 according to the patient's rehabilitation exercise index and physiological state to ensure the effectiveness and safety of the heat therapy. 3. Assisted flexion-extension mechanism 2: Assisted flexion-extension training: The assisted flexion and extension mechanism 2 is activated based on visual feedback and rehabilitation plan to assist the patient in flexion and extension training; the upper and lower leg straps fix the patient's thighs and calves respectively, and the knee strap is installed on the knee assist bracket to fix the patient's knee to the assist bracket; the main control chip controls the force change based on visual data, and the vertical rodless cylinder drives the knee assist bracket to move up and down to assist the patient in flexion and extension movements; 4. Evaluation and optimization of rehabilitation effects: Throughout the process, the GAN-generated model simulates the characteristics of standard rehabilitation movements, compares them with dynamic motion feature vectors, calculates similarity, and generates a rehabilitation movement index. If the rehabilitation movement index is lower than the threshold, the device automatically adjusts the resistance or assist parameters of the auxiliary flexion and extension mechanism 2 (such as reducing resistance to adapt to the patient's current ability), while enhancing the heat therapy intensity of key acupoints. 5. Remote monitoring and feedback: Doctors can view patient data through a terminal and adjust long-term training plans based on rehabilitation exercise indices; patients can receive real-time movement correction prompts through smart terminals to improve the standardization of training; rehabilitation data (such as movement trajectory, heat therapy temperature, and pain index) are uploaded to the cloud in real time, allowing doctors to remotely monitor and adjust the plan.
[0029] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. An intelligent rehabilitation device based on machine vision and thermotherapy technology, characterized in that, include: A multi-angle machine vision monitoring module is used to collect patients' lower limb movement data in real time and generate dynamic motion feature vectors; The intelligent positioning and temperature control module for the thermotherapy head includes a displacement adjustment mechanism, a thermotherapy head, and a temperature control system. The displacement adjustment mechanism adjusts the spatial position of the thermotherapy head according to the dynamic motion feature vector, and the temperature control system dynamically adjusts the thermotherapy temperature according to the patient's movement state. The assistive flexion and extension mechanism and linkage control module, including the knee assistive brace, strap and drive assembly, are used to assist patients in completing flexion and extension movements; The rehabilitation effect evaluation and parameter optimization module generates standard rehabilitation movement features based on generative adversarial networks, generates a rehabilitation movement index by comparing it with the dynamic movement feature vector, and optimizes the equipment operating parameters accordingly. The real-time feedback and remote monitoring module is used to provide real-time action correction prompts and remote data transmission; The multi-angle machine vision monitoring module, the intelligent positioning and temperature control module of the thermotherapy head, and the auxiliary flexion and extension mechanism and linkage control module achieve data fusion and closed-loop linkage control through the main control chip.
2. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, The multi-angle machine vision monitoring module includes three RGB-D cameras arranged in a 120° ring, installed at a height of 60-80cm above the ground. The external parameters are aligned using the Zhang Zhengyou calibration method, and a world coordinate system with the center of knee joint rotation as the origin is established.
3. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, The multi-angle machine vision monitoring module uses the HOG algorithm to extract limb contours and joint position features from keyframes, and performs multi-scale spatiotemporal feature fusion based on the Vision Transformer deep learning model to generate the dynamic motion feature vector.
4. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, The intelligent positioning and temperature control module of the thermotherapy head identifies the three-dimensional coordinates of the knee joint and key acupoints such as the inner knee eye and Yanglingquan based on visual data, and drives the displacement adjustment mechanism to automatically adjust the position of the thermotherapy head so that it is aligned with the target acupoints.
5. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, The displacement adjustment mechanism consists of a three-axis linear module and a two-degree-of-freedom rotary table. The three-axis linear module has a travel of ±200mm in the X, Y, and Z directions, and the two-degree-of-freedom rotary table can adjust the pitch and yaw by ±30°.
6. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, The assistive flexion and extension mechanism includes a knee assist bracket, upper and lower leg straps, a knee strap, and a vertical rodless cylinder. The piston rod of the rodless cylinder is mechanically connected to the knee assist bracket, and the main control chip controls the cylinder's extension and retraction speed and stroke based on visual data.
7. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, The main control chip integrates the motion trajectory from visual feedback, the temperature status of the thermotherapy head, and the preset rehabilitation plan to adjust the cylinder output force and motion parameters in real time. When the rehabilitation exercise index is lower than the set threshold, it automatically reduces resistance or provides appropriate assistance when the patient's movements are standardized.
8. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, The rehabilitation effect assessment and parameter optimization module simulates the standard rehabilitation movement characteristics under ideal conditions through a GAN generation model, generating standard fusion feature data covering normal gait cycle, joint angle variation range and limb coordination.
9. The intelligent rehabilitation device based on machine vision and thermotherapy technology according to claim 1, characterized in that, When the rehabilitation exercise index falls below the set threshold, the main control chip automatically triggers the parameter optimization process, dynamically adjusting the temperature distribution of the heat therapy head and the resistance or assist parameters of the flexion and extension mechanism to form a closed-loop control until the rehabilitation exercise index returns to the target range.
10. A method using the intelligent rehabilitation device based on machine vision and thermotherapy technology as described in any one of claims 1-9, characterized in that, Includes the following steps: The multi-angle machine vision monitoring module collects the patient's lower limb movement data in real time, and generates dynamic motion feature vectors by fusing the HOG algorithm with the Transformer model. Based on the dynamic motion feature vector, the coordinates of the knee joint and key acupoints are determined, the displacement adjustment mechanism is controlled to adjust the position of the heat therapy head, and the temperature control system is activated to adjust the temperature of the heat therapy head to the preset range. Based on the generative adversarial network, standard rehabilitation movement features are generated and compared with the dynamic movement feature vector to generate a rehabilitation movement index. The resistance / assistance parameters and thermotherapy temperature of the auxiliary flexion and extension mechanism are dynamically adjusted according to the rehabilitation exercise index to achieve synergistic optimization of rehabilitation training and thermotherapy.