Visual identification method for intelligent lifting physical therapy and intelligent lifting physical therapy equipment

Abstract

The invention relates to a visual identification method for intelligent pressing and lifting physiotherapy and intelligent pressing and lifting physiotherapy equipment, and the intelligent pressing and lifting physiotherapy equipment comprises a bed body which comprises a supporting surface, a lifting mechanism and a position adjusting device, the lifting mechanism is used for carrying out pressing and lifting operation, and the position adjusting device is used for adjusting the position of the lifting mechanism; the mechanical arm is arranged above the bed body and is used for carrying out pressing and lifting operation; the visual identification system comprises a camera arranged above the bed body, and the camera is used for acquiring body data of a person subjected to physical therapy; and the control system comprises a data processing system, the control system is used for controlling communication and operation of the bed body, the mechanical arm and the visual identification system, the data processing system receives the body data, and the lifting track is obtained by utilizing the visual identification method provided by the invention.

Description

Technical Field

[0001] The present application relates to the field of medical instruments, and in particular to a visual recognition method and intelligent massage therapy equipment for intelligent massage therapy. Background Art

[0002] Scoliosis is a common spinal disease characterized by lateral curvature of the spine in the coronal plane. Traditional treatments include surgery, braces, and physical therapy, but these methods often have certain limitations. In recent years, with the deepening of research on traditional Chinese medicine, Chinese medicine techniques such as massage have been found to have significant effects on improving scoliosis. Modern medical research shows that Chinese medicine guiding therapy improves the dynamic balance of the spine and restores bone straightness and tendon flexibility through breathing and breath exercises combined with limb stretching and pulling exercises.

[0003] However, traditional manual massage therapy relies on the doctor's experience and skills, and has problems such as irregular operation and improper strength control, and cannot last for a long time, which limits the treatment effect. On the other hand, there is a serious shortage of experienced medical resources in this field, and doctors are under heavy burden, so some patients in need cannot get effective treatment. Therefore, it is particularly important to develop an intelligent device that can accurately imitate the Chinese medicine massage technique and provide personalized treatment plans based on individual differences. Summary of the invention

[0004] The visual recognition method for intelligent massage therapy and the intelligent massage therapy device proposed in this application effectively solve the problems of irregular operation and inability to work for a long time in traditional manual massage therapy, and supplement a large amount of medical resources.

[0005] In a first aspect, the embodiments of the present application disclose a visual recognition method for intelligent massage therapy, comprising: Acquiring body data of a person being treated, wherein the body data includes image data and / or video data, or three-dimensional point cloud data; The key points of the body data are extracted using an artificial intelligence algorithm, wherein: When the body data includes image data and / or video data, detecting key points of the person being treated using a posture estimation algorithm; When the body data includes three-dimensional point cloud data, the method comprises the steps of: Preprocessing the three-dimensional point cloud data; Extracting the three-dimensional point cloud data belonging to the human body region of the person being treated from the three-dimensional point cloud data by using a region growing segmentation algorithm; Segmenting the human body region based on the three-dimensional point cloud data of the human body region; Acquire a skeleton model of the person being treated based on the segmented three-dimensional point cloud data of the human body region; Predicting the position of the joints of the person being treated by matching a predefined 3D human body model; Based on the position of the joint, using a pre-trained joint position regression model to obtain the key point; Based on the key points, a rocking trajectory is generated.

[0006] Optionally, the pose estimation algorithm includes an OpenPose algorithm.

[0007] Optionally, when the body data includes image data and / or video data, generating the rocking trajectory based on the key points includes: Based on the key points and the associated areas of the key points, obtaining the joint structure of the back of the person being treated; Based on the relative position and structure between the key points and the joint structure, calculating the back feature points of the person being treated, the back feature points including the position, size and distance data of the shoulders, spine and hips of the person being treated; Based on the back feature points, an initial rocking trajectory is generated by interpolation method; Smoothing the initial rocking trajectory; Based on the smoothed initial rocker pressing trajectory, a path planning algorithm is used to generate the rocker pressing trajectory.

[0008] Optionally, a Bezier curve is used to smooth the initial rocking trajectory.

[0009] Optionally, when the body data includes three-dimensional point cloud data, generating the rocking trajectory based on the key points includes: Based on the coordinates of the key points, mapping the joints into a coordinate system; Determining geometric information of the joint based on the position of the joint mapped into the coordinate system; Based on the geometric information of the joints, the motion trajectory of the person being treated is predicted through a Kalman filter algorithm, and a smooth massage trajectory is obtained.

[0010] In a second aspect, the embodiments of the present application disclose a smart massage therapy device, comprising: The bed body comprises a support surface, a lifting mechanism and a position adjustment device, wherein the lifting mechanism is used for performing a push-up operation, and the position adjustment device is used for adjusting the position of the lifting mechanism; A mechanical arm, arranged above the bed, for performing a step-pressing operation; A visual recognition system, comprising a camera disposed above the bed, the camera being used to obtain body data of the person being treated; A control system includes a data processing system, wherein the control system is used to control the communication and operation of the bed, the robotic arm and the visual recognition system, the data processing system receives the body data, and obtains the massage trajectory using any of the methods described above.

[0011] Optionally, the lifting mechanism includes an electric cylinder and a pressing head arranged on the top of the electric cylinder, the electric cylinder is used to lift or lower the height of the pressing head, and the pressing head is used to press the corresponding body part of the person being treated.

[0012] Optionally, the position adjustment device includes a uniaxial displacement mechanism or a synchronous displacement mechanism.

[0013] Optionally, the synchronous displacement mechanism comprises: Slider, connecting shaft and guide rail; wherein, The upper surface of the connecting shaft is provided with the lifting mechanism, and the lower surface is connected with the sliding block; The slider is used to slide on the guide rail to drive the connecting shaft to move, and a groove is provided at the bottom of the slider to limit the slider on the guide rail.

[0014] Optionally, the camera is a 2D camera or a 3D camera.

[0015] This application greatly improves the treatment efficiency and reduces the human involvement through the above-mentioned intelligent massage therapy device and the visual recognition method used by the intelligent massage therapy device, alleviates the serious shortage of experienced medical resources and the heavy burden on doctors in this field, and enables some patients who cannot be treated due to insufficient medical manpower to be effectively treated, and improves the accuracy and consistency of treatment. Furthermore, since the position and strength of the electric cylinder and the mechanical arm can be adjusted according to the specific situation of the patient, truly personalized treatment is achieved, avoiding the problem of unstable effects caused by individual differences in traditional manual treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a structural schematic diagram of the intelligent massage therapy device according to the implementation mode of the present application.

[0017] Figure 2 It is a structural schematic diagram of the bed body of the intelligent massage therapy device according to the implementation mode of the present application.

[0018] Figure 3 It is a structural schematic diagram of a single-axis displacement mechanism of an intelligent massage therapy device according to an embodiment of the present application.

[0019] Figure 4 It is a structural schematic diagram of the synchronous displacement mechanism of the intelligent massage therapy device according to the implementation mode of the present application.

[0020] Figure 5 It is a flow chart of a visual recognition method for intelligent massage therapy according to an embodiment of the present application.

[0021] Figure 6 It is a schematic diagram of performing a massage operation on a person being treated according to a generated massage trajectory according to an implementation mode of the present application. DETAILED DESCRIPTION

[0022] Figure 1 1 is a schematic diagram of the structure of the intelligent massage therapy device proposed in the present application. The intelligent massage therapy device comprises: a bed 1, a mechanical arm 2, a visual recognition system 3 and a control system 4.

[0023] The bed body 1 includes a support surface 11, a lifting mechanism 12 and a position adjustment device 13. The support surface 11 is used for the patient to lie face down on the support surface 11 for acupressure therapy; the lifting mechanism 12 can change the lifting height to perform acupressure operation on the patient lying on the support surface 11; the position adjustment device 13 is used to adjust the position of the lifting mechanism. For example, for patients with different physical conditions such as height, weight, gender, etc., the positions of various parts of their bodies on the support surface 11 are also different, so the position of the lifting mechanism 12 needs to be adjusted to correspond to the position of the corresponding body part of the patient to be acupressure.

[0024] The mechanical arm 2 is arranged above the bed body for performing the rocking operation. The lifting mechanism 12 in the bed body 1 simulates the rocking action in the rocking operation, and the mechanical arm 2 simulates the pressing action in the rocking operation, so that the lifting mechanism 12 and the mechanical arm 2 work together to perform a complete rocking operation.

[0025] The visual recognition system 3 includes a camera arranged above the bed 1, and the camera is used to photograph the person being treated to obtain the body data of the person being treated. After obtaining the body data, the obtained body data is processed by a data processing device to generate a massage trajectory suitable for the current person being treated. For different needs, the camera can be a 2D camera to obtain body data in the form of images or videos, or it can be a 3D camera to obtain body data in the form of three-dimensional point cloud data. The specific visual recognition method used by the visual recognition system 3 will be described in detail later. It can be understood that the data processing device can be integrated with the camera, and the data processing device can also be an external device. The camera transmits the body data to the data processing device for processing via wired, wireless or cloud. For example, the data processing device can also be integrated in the control system.

[0026] The control system 4 includes a data processing system. The control system 4 is mainly used to control the communication and operation of the bed 1, the mechanical arm 2 and the visual recognition system 3, and receives the body data collected by the camera through the data processing system, so as to process the body data using the visual recognition method, identify the body parts of the person being treated and the specific positions of each part, and then generate a massage trajectory suitable for the person being treated. The control system 4 is the control center of the entire intelligent massage therapy device, managing the information transmission, operation and drive between all components. In addition, the control system 4 can also perform personalized parameter adjustments, remotely monitor the massage therapy process, record various data during the massage therapy process, etc., and display it in real time through the human-computer interaction interface to facilitate doctors to track the progress of treatment and adjust the treatment plan in time. It is understandable that although Figure 1 The control system shown in the figure is a control cabinet connected to other components by wire, but the control system can be in various forms, such as a system connected to various components by wire, wireless, cloud, etc.

[0027] The operations of the lifting mechanism 12 and the position adjustment device 13 are controlled by the control system 4. Figure 2 - Figure 4 The structures and working principles of the lifting mechanism 12 and the position adjustment device 13 are described in detail.

[0028] like Figure 2 As shown in the structural diagram of the bed, the lifting mechanism 12 is arranged above the position adjustment device 13, so that the position of the lifting mechanism 12 in the bed can be adjusted by the position adjustment device 13. The lifting mechanism 12 may include an electric cylinder 121 and a pressing head 122 arranged on the top of the electric cylinder 121. Under the control of the control system 4, the electric cylinder 121 begins to adjust the height in sequence according to the pressing trajectory to adjust the height of the pressing head 122 to press the corresponding body part of the person being treated. It can be understood that the pressing head 122 can be designed into different shapes according to actual needs, and can be detachably arranged on the top of the lifting mechanism 12 for easy replacement.

[0029] The position adjustment device 13 may include a single-axis displacement mechanism or a synchronous displacement mechanism.

[0030] Figure 3 Schematic diagram of a uniaxial displacement mechanism, wherein each lifting mechanism 12 is connected to a uniaxial displacement platform module 31. Under the control of the control system 4, according to the position of each body part of the person being treated determined by the control system 4, each uniaxial displacement platform module can be independently displaced to drive the lifting mechanism 12 connected thereto to adjust the position. In one embodiment, seven groups of uniaxial displacement platform modules and corresponding lifting mechanisms can be set according to the parts to be treated, for example Figure 3In the a~g shown in the figure, a corresponds to the head of the person being treated, b corresponds to the left shoulder, c corresponds to the left waist, d corresponds to the left hip, e corresponds to the right shoulder, f corresponds to the right waist, and g corresponds to the right hip.

[0031] The synchronous displacement mechanism includes a connecting shaft 41, a slider 42 and a guide rail 43. The upper surface of the connecting shaft has one or more positions for setting the lifting mechanism, and the lower surface is connected to multiple sliders. The slider can slide on the guide rail to drive the upper connected connecting shaft to move, and then the positions of multiple lifting mechanisms 12 can be adjusted simultaneously. The groove set at the bottom of the slider limits the slider on the guide rail to prevent derailment.

[0032] like Figure 4 As shown, in one embodiment, three groups of synchronous displacement mechanisms and corresponding lifting mechanisms can be set according to the massaged parts: the first group includes a lifting mechanism A, corresponding to the head of the person being massaged, the lifting mechanism A is set on a connecting shaft, and slides on the guide rail through a slider to drive the lifting mechanism A to move; the second group includes a lifting mechanism B corresponding to the left shoulder, a lifting mechanism C corresponding to the left waist, and a lifting mechanism D corresponding to the left buttocks. The lifting mechanisms B, C, and D are connected together by a connecting shaft, and slide synchronously on three guide rails through three sliders respectively to drive the lifting mechanisms B, C, and D to move synchronously; the third group includes a lifting mechanism E corresponding to the right shoulder, a lifting mechanism F corresponding to the right waist, and a lifting mechanism G corresponding to the right buttocks. The lifting mechanisms E, F, and G are connected together by a connecting shaft, and slide synchronously on three guide rails through three sliders respectively to drive the lifting mechanisms E, F, and G to move synchronously. It is understandable that the lifting mechanisms A~G can also be arranged on the same connecting shaft, that is, the three connecting shafts shown in the figure are the same connecting shaft, so as to synchronously adjust the positions of all the lifting mechanisms; the multiple guide rails shown in the figure can also be connected together to form a guide rail. In addition, for each synchronous displacement mechanism, there can be a central slider, and the control system 4 controls a single or a small number of prime movers to generate power to move the central slider, so that the movement of the entire connecting shaft can be driven by the movement of the central slider.

[0033] The single-axis displacement mechanism can control each lifting mechanism individually, so the adjustment method is simple and the degree of freedom is high. However, the cost of the single-axis displacement platform module is high, which leads to a high cost for the entire mechanism. The synchronous displacement mechanism is low in cost because it only uses a mechanical structure, but it is difficult to adjust flexibly after the position of the guide rail is fixed, and it is impossible to achieve precise adjustment of each lifting mechanism separately. Different position adjustment devices can be selected or replaced according to needs.

[0034] Next reference Figure 5The flowchart shown is used to explain in detail the visual recognition method adopted by the visual recognition system 3.

[0035] In step 501, the body data of the person being treated is obtained. The body data is obtained by collecting the body data of the person being treated through the camera of the visual recognition system 3. The body data may be in different forms based on different cameras. For example, a 2D camera may obtain image data or video data, while a 3D camera may obtain three-dimensional point cloud data.

[0036] In step 502, the key points of the body data are extracted using an artificial intelligence algorithm. Based on the form of the body data collected in step 501, the key points of the body data can be extracted using different algorithms. This feature is based on the key points on the body of the person being treated that need to be massaged in the massage theory, such as shoulders, the middle point of the spine, and buttocks.

[0037] In step 503, a rocking trajectory is generated based on the key points. Based on the key points, the specific position of the body part of the person being treated on the bed and the height from the bed support surface and other data can be obtained, and then the rocking trajectory can be generated to determine the position where the lifting mechanism should be located and the height (and / or strength, depth, etc.) that the lifting mechanism and the mechanical arm need to press when performing the rocking operation, and the lifting mechanism and the mechanical arm are controlled by the control system to perform the rocking operation.

[0038] When the camera is a 2D camera, the collected body data can be in the form of image data and / or video data, and the processing methods for image data and video data are the same. After acquiring the body data in the form of image data and / or video data, the OpenPose algorithm is used to detect the key points of the person being treated, and the key points and the associated areas related to the key points can be vector-connected to further form the joint structure of the back of the human body. Through the joint structure and key point data of the back of the human body, the back feature points of the person being treated can be further obtained.

[0039] The back feature points are part of the key points, which are used to characterize the features of some key positions on the back of the human body, such as data reflecting the relative position and size between the key points on the back. In addition, the associated area refers to the local image area around the key points. These areas provide contextual information to help locate the key points more accurately, and can also be used to calculate the association relationship between key points, such as the connection parts of limbs. By analyzing the characteristics of the associated area, the direction and posture of the limbs can be better inferred, so as to accurately connect the key points to form a complete human posture estimation. In the algorithm, it can help determine whether different key points belong to the same human body. For example, by defining the associated area between the shoulder key point and the elbow key point, it can be made clear that these two points belong to the same arm. In other words, it is equivalent to connecting lines on the human skeleton structure. These lines and the areas surrounded by the lines are the associated areas, which ensure that the various parts will not be incorrectly assigned to different individuals during posture estimation.

[0040] Finally, based on the back feature points of the person being treated and the Chinese medicine massage therapy treatment plan, the initial massage trajectory suitable for the person being treated is obtained by interpolation, for example, by calculating the shortest path and curve between the back feature points (i.e., estimating the posture of the person being treated). In order to further obtain the optimal massage trajectory, the Bezier curve can be used to smooth the initial massage trajectory, and the path planning algorithm (such as Dijkstra, A*, etc.) can be used to further optimize the smoothed initial massage trajectory to obtain the optimal massage trajectory.

[0041] When the camera is a 3D camera, the collected body data can be in the form of three-dimensional point cloud data. After acquiring the body data in the form of three-dimensional point cloud data, the body data needs to be preprocessed first because the three-dimensional point cloud data usually contains noise, redundancy, unnecessary background information and other data that are irrelevant to the body of the person being treated. The preprocessing step may include: setting a threshold or depth range, eliminating data that exceeds the threshold or depth range to eliminate data that is not part of the human body, so as to remove background noise; using statistical filtering, Voxel Grid and other methods to reduce the noise of the point cloud and reduce the amount of data; uniformly sampling to reduce the number of data points, thereby improving data processing efficiency.

[0042] However, the preprocessed 3D point cloud data may still include data irrelevant to the body of the person being treated, and therefore the data needs to be further extracted, including the following steps: a segmentation algorithm based on region growing is used to separate and extract the data belonging to the human body region of the person being treated from the preprocessed 3D point cloud data according to the proximity relationship and curvature of each data point in the 3D point cloud data; a pre-trained 3D human body model is used, and based on the extracted 3D point cloud data of the human body region, segmentation is performed according to the human body region, for example, the 3D point cloud data can be segmented into shoulder region, waist region, head region, hip region and other regions, to facilitate subsequent data processing.

[0043] Next, the key points of the body of the person being treated can be extracted based on the three-dimensional point cloud data segmented by human body regions, including the steps of: skeletonizing the three-dimensional point cloud data segmented by human body regions to obtain a skeleton model of the body of the person being treated; matching the skeleton model with a predefined 3D human body model to predict the positions of the various joints of the person being treated. The three-dimensional point cloud data is further divided into multiple data blocks according to the joint positions, and the joint positions are regressed to the real body of the person being treated through a joint position regression model obtained by pre-training a convolutional neural network to obtain the key points of the body of the person being treated.

[0044] After obtaining the key points of the patient's body, the joints are mapped to the coordinate system based on the coordinates of the key points by combining reprojection technology, so that the geometric information of each joint can be obtained, and the angle of the joint can be further calculated. Finally, the Kalman filter algorithm is used to predict the patient's motion trajectory by combining the angle of the joint and other information, as well as the point cloud data and prior information of the continuous frames, and finally a smooth massage trajectory is generated.

[0045] Figure 6 This is a non-limiting example of performing a rocking operation based on the generated rocking trajectory. Press 1 (left shoulder), press 2 (left waist), press 3 (right waist) are the operations performed by the robot arm on the back of the person being treated, and rocking 1 (head), rocking 2 (right shoulder), rocking 3 (left waist), and rocking 4 (right hip) are the operations performed by the lifting mechanism on the front of the person being treated. The rocking trajectory is that, in the first step, press 1 and rocking 1 are performed synchronously; in the second step, press 2 and rocking 2 are performed synchronously; in the third step, press 3 and rocking 3 are performed synchronously; in the fourth step, rocking 4 is performed. It can be understood that this scheme is only used as an example, and different rocking trajectories will be generated for different people being treated in practice.

[0046] The above specific embodiments illustrate the implementation of the present application, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Although the description of the present application will be introduced in conjunction with the preferred embodiment, this does not mean that the features of this invention are limited to the implementation. In addition, in order to avoid confusion or blurring the focus of the present application, some specific details will be omitted in the description. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0047] Furthermore, various operations will be described as multiple discrete operations in a manner that is most helpful for understanding the illustrative embodiments; however, the order of description should not be construed as implying that these operations are necessarily order dependent. In particular, these operations do not need to be performed in the order presented.

[0048] Unless the context dictates otherwise, the terms "comprising," "having," and "including" are synonymous. The phrase "A / B" means "A or B." The phrase "A and / or B" means "(A and B) or (A or B)."

[0049] As used herein, the term "module" or "unit" may refer to, be or include: an application specific integrated circuit (ASIC), an electronic circuit, a (shared, dedicated or group) processor and / or memory that executes one or more software or firmware programs, a combinational logic circuit and / or other suitable components that provide the described functionality.

[0050] In the accompanying drawings, some structural or method features are shown in a specific arrangement and / or order. However, it should be understood that such a specific arrangement and / or order may not be required. In some embodiments, these features may be arranged in a manner and / or order different from that shown in the illustrative drawings. In addition, the inclusion of structural or method features in a specific figure does not mean that such features are required in all embodiments, and in some embodiments, these features may not be included or may be combined with other features.

[0051] It should be understood that although the terms "first", "second", etc. may be used here to describe various units or data, these units or data should not be limited by these terms. These terms are used only to distinguish one feature from another. For example, without departing from the scope of the exemplary embodiment, a first feature may be referred to as a second feature, and similarly, a second feature may be referred to as a first feature.

[0052] It should be noted that in this specification, similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0053] While the present invention has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.

Claims

1. A visual recognition method for intelligent massage therapy, characterized in that: include: Acquiring body data of a person being treated, wherein the body data includes image data and / or video data, or three-dimensional point cloud data; The key points of the body data are extracted using an artificial intelligence algorithm, wherein: When the body data includes image data and / or video data, detecting key points of the person being treated using a posture estimation algorithm; When the body data includes three-dimensional point cloud data, the method comprises the steps of: Preprocessing the three-dimensional point cloud data; Extracting the three-dimensional point cloud data belonging to the human body region of the person being treated from the three-dimensional point cloud data by using a region growing segmentation algorithm; Segmenting the human body region based on the three-dimensional point cloud data of the human body region; Acquire a skeleton model of the person being treated based on the segmented three-dimensional point cloud data of the human body region; Predicting the position of the joints of the person being treated by matching a predefined 3D human body model; Based on the position of the joint, using a pre-trained joint position regression model to obtain the key point; Based on the key points, a rocking trajectory is generated.

2. The method according to claim 1, characterized in that The pose estimation algorithm includes an OpenPose algorithm.

3. The method according to claim 1, characterized in that When the body data includes image data and / or video data, generating the rocking trajectory based on the key points includes: Based on the key points and the associated areas of the key points, obtaining the joint structure of the back of the person being treated; Based on the relative position and structure between the key points and the joint structure, calculating the back feature points of the person being treated, the back feature points including the position, size and distance data of the shoulders, spine and hips of the person being treated; Based on the back feature points, an initial rocking trajectory is generated by interpolation method; Smoothing the initial rocking trajectory; Based on the smoothed initial rocker pressing trajectory, a path planning algorithm is used to generate the rocker pressing trajectory.

4. The method according to claim 3, characterized in that: The initial rocker trajectory is smoothed using a Bezier curve.

5. The method according to claim 1, characterized in that: When the body data includes three-dimensional point cloud data, generating the rocking trajectory based on the key points includes: Based on the coordinates of the key points, mapping the joints into a coordinate system; Determining geometric information of the joint based on the position of the joint mapped into the coordinate system; Based on the geometric information of the joints, the motion trajectory of the person being treated is predicted through a Kalman filter algorithm, and a smooth massage trajectory is obtained.

6. An intelligent massage therapy device, characterized in that: include: The bed body comprises a support surface, a lifting mechanism and a position adjustment device, wherein the lifting mechanism is used for performing a push-up operation, and the position adjustment device is used for adjusting the position of the lifting mechanism; A mechanical arm, arranged above the bed, for performing a step-pressing operation; A visual recognition system, comprising a camera disposed above the bed, the camera being used to obtain body data of the person being treated; A control system, including a data processing system, wherein the control system is used to control the communication and operation of the bed, the robotic arm and the visual recognition system, the data processing system receives the body data, and obtains the massage trajectory using the method described in any one of claims 1-5.

7. The device according to claim 6, characterized in that The lifting mechanism includes an electric cylinder and a pressing head arranged on the top of the electric cylinder, the electric cylinder is used to lift or lower the height of the pressing head, and the pressing head is used to press the corresponding body part of the person being treated.

8. The device according to claim 7, characterized in that The position adjustment device includes a single-axis displacement mechanism or a synchronous displacement mechanism.

9. The device according to claim 8, characterized in that The synchronous displacement mechanism comprises: Slider, connecting shaft and guide rail; wherein, The upper surface of the connecting shaft is provided with the lifting mechanism, and the lower surface is connected with the sliding block; The slider is used to slide on the guide rail to drive the connecting shaft to move, and a groove is provided at the bottom of the slider to limit the slider on the guide rail.

10. The device according to claim 9, characterized in that The camera is a 2D camera or a 3D camera.

Images