Infrared physiotherapy visual guidance training method and system based on VR interaction

Abstract

The invention provides an infrared physiotherapy visual guidance training method and system based on VR interaction, and the method comprises the steps: firstly obtaining the body surface temperature distribution data of a user, and generating an infrared thermogram containing temperature gradient layering and low-temperature sensitive region marking; generating a three-dimensional physiotherapy guide model containing a guide action path and a thermal image feedback area by combining historical interaction action trajectory data, presenting a dynamic guide interface through VR display equipment, and adjusting action prompt intensity according to a deviation value between real-time interaction action trajectory data and the guide action path. And training effect feedback information is generated according to the matching degree and the difference value, so that visual and intelligent guidance of infrared physiotherapy training is realized, a personalized training scheme can be provided according to the actual situation of the user, the training effect and efficiency are improved, and personalized requirements are met.

Description

Technical Field

[0001] This invention relates to the field of intelligent medical hardware interaction technology based on virtual reality, and more specifically, to a VR-interactive infrared physiotherapy visualization-guided training method and system. Background Technology

[0002] In traditional infrared therapy training, users often find it difficult to intuitively understand the actual condition of their treatment area and the relationship between the therapy movements and the therapeutic effect. Existing infrared therapy methods typically only use infrared devices for irradiation, and users cannot clearly see the specific temperature distribution of the treatment area, lacking intuitive understanding of key information such as low-temperature sensitive areas.

[0003] Meanwhile, in guiding physical therapy exercises, traditional methods mainly rely on verbal instructions from therapists or simple illustrated explanations. This approach is not only insufficiently intuitive but also lacks the ability to provide real-time, precise feedback and adjustments based on the user's actual movements. Users often struggle to accurately grasp the correct form and effectiveness of their movements during physical therapy, leading to poor therapeutic results. Furthermore, the lack of comprehensive analysis and utilization of users' historical physical therapy training data makes it impossible to provide personalized, continuously optimized physical therapy training programs, thus failing to meet the individualized needs of different users. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a VR-interactive infrared physiotherapy visualization-guided training method, the method comprising: The system acquires the user's current body surface temperature distribution data and the historical interaction trajectory data recorded by the VR device. The body surface temperature distribution data is acquired through an infrared sensing module, and the historical interaction trajectory data includes the user's limb spatial coordinate sequence at each time point in past physical therapy training. An infrared thermal image covering the target physiotherapy area is generated based on the body surface temperature distribution data. The infrared thermal image contains temperature gradient stratification information and low-temperature sensitive area markings based on a preset threshold. The historical interactive action trajectory data and the infrared thermal image are input into the physiotherapy model construction module. Through spatial location correlation analysis, a three-dimensional physiotherapy guidance model containing the guided action path and thermal image feedback area is generated. In the three-dimensional physiotherapy guidance model, each guided action node corresponds to the expected temperature change value of a thermal image feedback area. The dynamic guidance interface of the three-dimensional physiotherapy guidance model is presented to the user through a VR display device. Real-time interactive action trajectory data of the user is collected simultaneously. The intensity of the action prompts of the dynamic guidance interface is adjusted based on the deviation between the real-time interactive action trajectory data and the guidance action path. Based on the matching degree between the real-time interactive motion trajectory data and the guided motion path, and the difference between the actual temperature change value and the expected temperature change value of the thermal imaging feedback area, training effect feedback information containing motion adjustment suggestions and thermal imaging optimization goals is generated.

[0005] In another aspect, embodiments of the present invention also provide an infrared physiotherapy visualization-guided training system based on VR interaction, including a processor and a machine-readable storage medium. The machine-readable storage medium is connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code. The processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0006] Based on the above, by acquiring the user's current body surface temperature distribution data, an infrared thermogram containing temperature gradient stratification information and low-temperature sensitive area markers is generated, allowing the user to intuitively understand the temperature status of the physiotherapy area. Historical interactive action trajectory data and the infrared thermogram are input into the physiotherapy model building module to generate a 3D physiotherapy guidance model, realizing the correlation between physiotherapy actions and temperature changes in the physiotherapy area, providing users with precise action guidance. A dynamic guidance interface is presented through a VR display device, and the intensity of action prompts is adjusted based on the deviation between real-time interactive action trajectory data and the guided action path, enabling real-time and dynamic guidance for users to perform correct physiotherapy actions. Finally, training effect feedback information is generated based on the matching degree between real-time interactive action trajectory data and the guided action path, as well as the difference between the actual and expected temperature changes in the thermal image feedback area. This provides users with comprehensive training effect evaluation and optimization suggestions, enabling personalized training programs based on the user's actual situation, significantly improving the effectiveness and efficiency of infrared physiotherapy training, and meeting the user's personalized needs. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of the execution flow of the VR-interactive infrared physiotherapy visualization-guided training method provided in an embodiment of the present invention.

[0008] Figure 2 This is a schematic diagram of exemplary hardware and software components of the VR-interactive infrared physiotherapy visualization-guided training system provided in an embodiment of the present invention. Detailed Implementation

[0009] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a VR-interactive-based infrared physiotherapy visualization-guided training method according to an embodiment of the present invention. The following is a detailed description of this VR-interactive-based infrared physiotherapy visualization-guided training method.

[0010] Step S110: Obtain the user's current body surface temperature distribution data and the historical interaction trajectory data recorded by the VR device. The body surface temperature distribution data is acquired through an infrared sensing module, and the historical interaction trajectory data includes the user's limb spatial coordinate sequence at each time point in past physical therapy training.

[0011] In this embodiment, the body surface temperature distribution data reflects the current temperature of the user's body surface and is collected through an infrared sensing module. Historical interaction motion trajectory data records the user's movements during past physiotherapy training, including the spatial coordinate sequence of limbs at each time point; this data is recorded by the VR device. For example, in a real-world physiotherapy scenario, when a user enters a physiotherapy environment equipped with an infrared sensing module and a VR device, the system will begin the data acquisition process.

[0012] Step S111: Activate the multi-array infrared sensor set in the physiotherapy chamber to continuously scan the temperature of the user's target physiotherapy area and obtain multiple frames of raw temperature data with timestamps.

[0013] In this step, the multi-array infrared sensor installed in the physiotherapy chamber is activated. The advantage of the multi-array infrared sensor is its ability to scan the user's target physiotherapy area from multiple angles and positions, thereby obtaining more comprehensive and accurate temperature information. The multi-array infrared sensor can continuously scan the target physiotherapy area at preset time intervals, obtaining a frame of raw temperature data with each scan. Furthermore, to facilitate subsequent temporal analysis and processing of the data, each frame of raw temperature data is timestamped. Taking the user's lumbar physiotherapy as an example, the multi-array infrared sensor will continuously scan the lumbar region, recording the temperature data of that area at regular intervals, forming a series of raw temperature data frames with timestamps. These data frames contain temperature information for various locations on the lumbar region at different times.

[0014] Step S112: Perform spatial coordinate calibration on the raw temperature data, mapping the sensor coordinates of each frame of raw temperature data to the user's body surface anatomical coordinate system to generate body surface temperature distribution data.

[0015] Since the raw temperature data is collected based on the sensor coordinate system, it needs to be correlated with the user's actual body surface location. Therefore, spatial coordinate calibration is required to map the sensor coordinates to the user's body surface anatomical coordinate system.

[0016] Step S1121: Obtain three-dimensional scan data of the user's body surface anatomical structure and construct a body surface coordinate system containing key anatomical landmarks.

[0017] In this embodiment, to construct the user's body surface coordinate system, it is first necessary to acquire three-dimensional scan data of the user's anatomical structure. This can be accomplished, for example, using a high-precision three-dimensional scanning device, such as a laser three-dimensional scanner. The three-dimensional scanning device will perform a comprehensive scan of the user's body, acquiring the three-dimensional coordinate information of various points on the body surface. After obtaining the three-dimensional scan data, it is necessary to identify key anatomical landmarks. These key anatomical landmarks are usually distinctive and relatively fixed locations on the body, such as the acromion of the shoulder, the lateral epicondyle of the humerus at the elbow, and the styloid process of the ulna at the wrist. Based on these key anatomical landmarks, a coordinate system that can accurately describe the user's body surface position is constructed. For example, with the acromion of the shoulder as the origin, the direction connecting the left and right acromions as the X-axis, the direction perpendicular to the front of the body as the Y-axis, and the direction perpendicular to the ground upwards as the Z-axis, a three-dimensional body surface coordinate system is constructed.

[0018] Step S1122: Extract temperature sampling points from the original temperature data that correspond to the spatial locations of key anatomical landmarks.

[0019] After constructing the body surface coordinate system, it is necessary to extract temperature sampling points corresponding to the spatial locations of key anatomical landmarks from the raw temperature data. This requires preliminary mapping and matching between the sensor coordinates and the body surface coordinate system. First, based on the coordinates of the key anatomical landmarks in the body surface coordinate system, find the temperature sampling point in the raw temperature data that has the closest spatial location. Since there may be some deviation between the sensor coordinates and the coordinates in the body surface coordinate system, precise matching is required. For example, a nearest neighbor algorithm can be used to calculate the distance between the coordinates of each key anatomical landmark in the body surface coordinate system and the sensor coordinates of each temperature sampling point in the raw temperature data, and select the closest temperature sampling point as the temperature sampling point corresponding to that key anatomical landmark.

[0020] Step S1123: Calculate the coordinate transformation matrix between the sensor coordinates of the temperature sampling point and the coordinates of the corresponding anatomical landmark in the body surface coordinate system. The coordinate transformation matrix includes a translation vector and a rotation matrix.

[0021] After obtaining the temperature sampling points corresponding to key anatomical landmarks, the next step is to calculate the coordinate transformation matrix. This matrix maps the sensor coordinates to the body surface coordinate system. The coordinate transformation matrix contains translation vectors and rotation matrices. The translation vectors describe the relative translation relationship between the two coordinate systems in space, and the rotation matrix describes the relative rotation relationship between the two coordinate systems. The specific calculation process is as follows: First, select at least three non-collinear key anatomical landmarks and their corresponding temperature sampling points. Let the coordinates of these key anatomical landmarks in the body surface coordinate system be P1, P2, and P3, and the coordinates of the corresponding temperature sampling points in the sensor coordinate system be Q1, Q2, and Q3.

[0022] Then, using methods such as the least squares method, the translation vector T and rotation matrix R are solved to ensure that, after coordinate transformation, the coordinates of the temperature sampling points are as close as possible to the coordinates of the corresponding key anatomical landmarks. Specifically, the following relationship needs to be satisfied: Qi = R * Pi + T (i = 1, 2, 3). By solving this system of equations, the translation vector T and rotation matrix R can be obtained, thus yielding the coordinate transformation matrix.

[0023] Step S1124: Perform coordinate transformation processing on all sampling points of the original temperature data through a coordinate transformation matrix to generate body surface temperature distribution data mapped to the body surface anatomical coordinate system.

[0024] After obtaining the coordinate transformation matrix, it is applied to all sampling points of the original temperature data. For each sampling point in the original temperature data, the coordinate transformation matrix is ​​used to transform it from sensor coordinates to the body surface anatomical coordinate system. By performing the above coordinate transformation process on all sampling points, body surface temperature distribution data mapped to the body surface anatomical coordinate system can be obtained. This body surface temperature distribution data can accurately reflect the temperature at various locations on the user's body surface.

[0025] Step S113: Retrieve the user's historical physiotherapy training log stored in the VR device, and extract the action recording segments that contain the spatial coordinates of limb joints. The time span of the action recording segments is the most recent preset training cycle.

[0026] After obtaining the user's current body surface temperature distribution data, it is also necessary to obtain the user's historical interaction movement trajectory data. This step is achieved by retrieving the user's historical physiotherapy training logs stored in the VR device. During past physiotherapy training, the VR device can record various movement information of the user, including the spatial coordinates of limb joints. Movement record segments containing the spatial coordinates of limb joints are extracted from the historical physiotherapy training logs, and the time span of these movement record segments is limited to the most recent preset training period. For example, if the preset training period is the most recent month, then the spatial coordinate records of the user's limb joints at various time points during physiotherapy training within the most recent month are extracted from the VR device's logs.

[0027] Step S114: Perform time alignment processing on the action recording segments to generate historical interactive action trajectory data containing continuous time series. The time resolution of the historical interactive action trajectory data is consistent with the scanning frequency of the body surface temperature distribution data.

[0028] After extracting the motion recording segments, time alignment is necessary because these segments may exhibit temporal discontinuities or inconsistent time intervals. The purpose of time alignment is to ensure that the historical interactive motion trajectory data has a continuous time series, and that its temporal resolution matches the scanning frequency of the body surface temperature distribution data. This ensures that the body surface temperature distribution data and the historical interactive motion trajectory data accurately correspond in time during subsequent data analysis and processing. Specifically, time alignment can be achieved through methods such as interpolation and resampling. For example, if the scanning frequency of the body surface temperature distribution data is once per second, but the time intervals of the motion recording segments are inconsistent, then methods such as linear interpolation can be used to insert data from missing time points into the motion recording segments, making the time intervals of the motion recording segments also once per second, thereby generating historical interactive motion trajectory data containing a continuous time series.

[0029] Step S120: Generate an infrared thermal image covering the target physiotherapy area based on the body surface temperature distribution data. The infrared thermal image contains temperature gradient stratification information and low-temperature sensitive area markings based on preset thresholds.

[0030] After obtaining the user's current body surface temperature distribution data, the next step is to generate an infrared thermogram covering the target treatment area based on this data. The infrared thermogram can visually display the temperature distribution of the user's body surface and includes temperature gradient stratification information and markings of low-temperature sensitive areas.

[0031] Step S121: Perform interpolation to complete the body surface temperature distribution data, fill in the temperature values ​​in the missing data areas, and generate continuous temperature field data that fully covers the target physiotherapy area.

[0032] When collecting body surface temperature distribution data, some areas may be missing data, for example, due to sensor obstruction or other reasons that prevent temperature data from being collected at certain locations. Therefore, interpolation is needed to complete the body surface temperature distribution data and fill in the temperature values ​​in these missing areas.

[0033] Step S1211: Identify missing regions in the body surface temperature distribution data. Missing regions are sets of continuous spatial points whose temperature values ​​are marked as invalid.

[0034] First, it's necessary to identify missing regions in the body surface temperature distribution data. These missing regions are typically continuous sets of spatial points where the temperature value is marked as invalid. For example, in body surface temperature distribution data, some areas might have temperature values ​​marked as invalid (such as NaN), and these areas are considered missing regions. The missing regions can be identified by iterating through the body surface temperature distribution data, finding all points with invalid temperature values, and merging adjacent invalid value points into continuous regions.

[0035] Step S1212: Extract the effective temperature values ​​of the missing region boundary and construct a temperature prediction function based on bicubic spline interpolation.

[0036] After identifying the missing region, the effective temperature values ​​of its boundaries are extracted. These boundary values ​​can serve as reference points for interpolation. Based on these boundary values, a temperature prediction function based on bicubic spline interpolation is constructed. Bicubic spline interpolation is a commonly used interpolation method that can construct a smooth interpolation function on a two-dimensional plane based on known discrete point values. Specifically, by analyzing and processing the effective temperature values ​​of the missing region boundaries, a function capable of predicting the temperature within the missing region is constructed using the bicubic spline interpolation algorithm. This temperature prediction function can calculate the predicted temperature value at the location within the missing region based on its spatial coordinates.

[0037] Step S1213: Input the spatial coordinates of the missing region into the temperature prediction function to calculate the predicted temperature value of the missing region.

[0038] After obtaining the temperature prediction function, the spatial coordinates of the missing region are input into the function. For each spatial point within the missing region, the predicted temperature value of that point is calculated using the temperature prediction function based on its coordinates. For example, for a spatial point (x, y) within the missing region, its coordinates are substituted into the temperature prediction function, and the predicted temperature value of that point is obtained through the calculation by the temperature prediction function.

[0039] Step S1214: Fill the missing regions with the predicted temperature values ​​to generate continuous temperature field data.

[0040] The calculated predicted temperature values ​​are then filled into the missing regions. For each spatial point within a missing region, its predicted temperature value replaces the original invalid temperature value. Through this filling operation, continuous temperature field data that fully covers the target treatment area can be obtained. This data no longer has any missing data and can more accurately reflect the temperature distribution of the target treatment area.

[0041] Step S122: Calculate the temperature change rate of adjacent spatial points based on the continuous temperature field data, and construct a temperature gradient field model based on the temperature change rate.

[0042] After obtaining continuous temperature field data, the rate of temperature change at adjacent spatial points is further calculated. The rate of temperature change reflects the spatial variation of temperature and is crucial for understanding the characteristics of temperature distribution. For each spatial point in the continuous temperature field data, the temperature difference between it and its adjacent spatial points is calculated and divided by the distance between the two points to obtain the rate of temperature change at that point in each direction. For example, for a spatial point P, its adjacent spatial points are P1, P2, P3, etc. The temperature difference ΔT1 between P and P1, and the distance d1 between P and P1, are calculated. Then, the rate of temperature change of P in the direction of P1 is r1 = ΔT1 / d1. By performing the above calculations for all spatial points in the continuous temperature field data, the rate of temperature change at each spatial point in each direction can be obtained.

[0043] Based on these rates of temperature change, a temperature gradient field model is constructed. This model describes the trend and intensity of temperature change in space. In the temperature gradient field model, each spatial point has a corresponding temperature gradient vector. The direction of this vector indicates the direction of the fastest temperature change, and the magnitude of the vector represents the intensity of the temperature change. By further processing and analyzing the rates of temperature change, such as calculating the magnitude and direction of the gradient vector, the temperature gradient field model can be constructed. This model can intuitively demonstrate the changes in temperature in space.

[0044] Step S123: According to the preset temperature range division rules, the continuous temperature field data is mapped to the thermal image color coding table to generate temperature gradient layering information with layered color identification.

[0045] After obtaining the continuous temperature field data and temperature gradient field model, the continuous temperature field data needs to be visualized to more intuitively display the temperature distribution. This step is achieved by mapping the continuous temperature field data to a thermal imaging color coding table according to a preset temperature range division rule.

[0046] The preset temperature range division rules are set based on actual needs and experience. For example, the temperature range can be divided into multiple intervals, each corresponding to a different color. The thermal imaging color coding table is a predefined table that records the color corresponding to each temperature interval. For each spatial point in the continuous temperature field data, the corresponding temperature interval is looked up based on its temperature value, and the color corresponding to that interval is obtained from the thermal imaging color coding table. Through the above mapping operation, the continuous temperature field data is converted into temperature gradient layering information with layered color labels. This information is displayed in the form of an image, with different colors representing different temperature intervals, thus allowing a visual understanding of the temperature distribution and gradient.

[0047] Step S124: Extract the boundaries of regions with temperature values ​​below a first preset threshold from the continuous temperature field data, connect the discrete low-temperature regions through morphological closing operations, and generate low-temperature sensitive region markers. The low-temperature sensitive region markers include the region contour coordinate sequence and the average temperature value within the region.

[0048] After generating the temperature gradient stratification information, it is also necessary to identify low-temperature sensitive areas. This is achieved by extracting the boundaries of regions in the continuous temperature field data whose temperature values ​​are below a first preset threshold. The first preset threshold is set based on actual physiotherapy needs and experience; regions below this threshold are considered low-temperature sensitive areas. By traversing the continuous temperature field data, all points with temperature values ​​below the first preset threshold are identified, and the boundaries of the regions formed by these points are determined.

[0049] However, these low-temperature regions may be discrete. To more accurately identify and process these regions, morphological closing operations are needed to connect them. Morphological closing is an image processing method that first performs a dilation operation to connect discrete regions, and then performs an erosion operation to restore the shape of the regions. Through morphological closing, discrete low-temperature regions are connected into one or more continuous regions.

[0050] Low-temperature sensitive area markers are generated, containing a sequence of region contour coordinates and an average temperature value within the region. For connected low-temperature regions, their contour coordinate sequences are extracted, accurately describing the region's shape and location. Simultaneously, the average temperature value of all points within the low-temperature region is calculated as the average temperature value for that region.

[0051] Step S130: Input the historical interactive action trajectory data and infrared thermal image into the physiotherapy model construction module, and generate a three-dimensional physiotherapy guidance model containing the guided action path and thermal image feedback area through spatial location correlation analysis. In the three-dimensional physiotherapy guidance model, each guided action node corresponds to the expected temperature change value of a thermal image feedback area.

[0052] After acquiring the user's body surface temperature distribution data and generating an infrared thermogram, as well as historical interaction trajectory data, these two data points are input into the physiotherapy model construction module. Spatial location correlation analysis is then used to generate a three-dimensional physiotherapy guidance model. This three-dimensional physiotherapy guidance model will provide users with specific physiotherapy guidance information.

[0053] Step S131: Perform principal component analysis on the historical interaction action trajectory data to extract the user's high-frequency repeated action patterns as basic action units.

[0054] First, principal component analysis (PCA) is performed on the historical interaction motion trajectory data. PCA is a method for data dimensionality reduction and feature extraction, which can identify the most important features and patterns in the data. Historical interaction motion trajectory data contains various movement information from the user's past physical therapy training. Through PCA, this movement information is dimensionality-reduced to identify the principal components, which represent the user's frequently repeated movement patterns and are used as basic movement units. For example, in the user's historical interaction motion trajectory data, there may be some frequently repeated movement patterns such as shoulder rotation and elbow flexion / extension. These movement patterns are extracted through PCA and used as basic movement units.

[0055] Step S132: Calculate the spatial overlap between the limb spatial coordinates of the basic motion unit and the contour coordinates of the low-temperature sensitive area of ​​the infrared thermogram, and select basic motion units whose overlap with the low-temperature sensitive area exceeds the second preset threshold as candidate guided motions.

[0056] After obtaining the basic motion unit, the spatial overlap between its limb spatial coordinates and the contour coordinates of the low-temperature sensitive area in the infrared thermogram is calculated. The spatial overlap calculation measures the degree of overlap between the limb spatial range of the basic motion unit and the low-temperature sensitive area. For each basic motion unit, the overlap area between the area covered by its limb spatial coordinates and the area defined by the contour coordinates of the low-temperature sensitive area is calculated, and this overlap area is compared with the limb spatial area of ​​the basic motion unit to obtain the spatial overlap.

[0057] Basic movement units whose overlap with the hypothermia-sensitive area exceeds a second preset threshold are selected as candidate guided movements. The second preset threshold is set based on actual physiotherapy needs and experience. Only when the overlap between the basic movement unit and the hypothermia-sensitive area exceeds this threshold is the basic movement unit considered potentially helpful for the physiotherapy of the hypothermia-sensitive area and thus selected as a candidate guided movement. For example, if the limb spatial range of a basic movement unit overlaps with the hypothermia-sensitive area by 80%, and the second preset threshold is 70%, then this basic movement unit will be selected as a candidate guided movement.

[0058] Step S133: Reassemble the candidate guided actions in time series to generate a guided action path with continuous motion logic. The guided action path includes the limb target coordinates and action speed constraints at each time node.

[0059] After obtaining candidate guiding actions, they need to be reorganized in time sequence to generate guiding action paths with continuous motion logic.

[0060] Step S1331: Extract the start and end coordinates of the candidate guiding actions and construct the motion direction vector between action nodes.

[0061] First, extract the start and end coordinates of the candidate guided actions. For each candidate guided action, its limb space coordinates at the start and end are used as the start and end coordinates, respectively. Based on the start and end coordinates, construct the motion direction vector between action nodes. The motion direction vector represents the direction of movement from the start point to the end point. For example, for a candidate guided action with start coordinate A and end coordinate B, the motion direction vector V can be calculated as follows: V = BA. By performing the above processing on all candidate guided actions, the motion direction vector between each action node is obtained.

[0062] Step S1332: Calculate the spatial distance between the termination coordinates of adjacent candidate guide actions and the starting coordinates of the next candidate guide action, and filter candidate guide action pairs whose spatial distance is less than the fourth preset threshold.

[0063] After obtaining the motion direction vector of each candidate guiding action, it is necessary to further analyze the connection relationship between adjacent candidate guiding actions. The spatial distance between the termination coordinates of adjacent candidate guiding actions and the starting coordinates of the next candidate guiding action is calculated. This spatial distance reflects the spatial proximity of the two adjacent actions. For each pair of adjacent candidate guiding actions, the spatial distance is obtained by calculating the magnitude of the difference between their termination and starting coordinates. For example, given adjacent candidate guiding actions C1 and C2, where C1's termination coordinate is E1 and C2's starting coordinate is S2, the spatial distance D between them can be obtained by calculating the magnitude of the difference between E1 and S2. Candidate guiding action pairs with a spatial distance less than a fourth preset threshold are selected. This fourth preset threshold is set based on the actual requirements for motion rationality and continuity. Only when the spatial distance between adjacent candidate guiding actions is less than this threshold are the two actions considered to be able to connect relatively naturally. These candidate guiding action pairs are then selected for subsequent time interval adjustment and trajectory fitting.

[0064] Step S1333: Adjust the time interval of the candidate guiding action pair so that the end time of the previous action is continuously connected with the start time of the next action.

[0065] For the selected candidate guided action pairs, the time interval needs to be adjusted. Since the candidate guided actions are extracted from historical interaction action trajectory data, they may not be sequentially connected in time. To generate guided action paths with continuous motion logic, the time intervals of these candidate guided action pairs need to be adjusted so that the end time of the previous action is continuously connected to the start time of the next action. This may involve modifying and rearranging the timestamps of the actions. For example, if the end time of the previous candidate guided action C1 is T1, and the start time of the next candidate guided action C2 is T2, and T2 is greater than T1 by a certain time interval, then the start time of C2 needs to be adjusted to T1, and the timestamps of other time nodes in C2 should be adjusted accordingly to ensure that the overall time sequence and duration of the actions are reasonable.

[0066] Step S1334: Fit the motion trajectory of the candidate guided action pair using Bézier curves to generate a smooth transition continuous motion path segment.

[0067] After adjusting the time intervals of the candidate guided action pairs, a Bézier curve is used to fit the motion trajectory of the candidate guided action pairs to make the transition between actions smoother and more natural. A Bézier curve is a mathematical tool that can generate a smooth curve based on given control points. For each candidate guided action pair, their start coordinates, end coordinates, and some key intermediate coordinate points are used as control points for the Bézier curve. Using a Bézier curve generation algorithm, a smooth curve is generated based on these control points. This curve represents the smooth transition trajectory between the two candidate guided actions. For example, for candidate guided action pairs C1 and C2, the end coordinate of C1, the start coordinate of C2, and some intermediate points selected between these two coordinates are used as control points to generate a Bézier curve. The curve generated in this way ensures a smooth transition from C1 to C2, avoiding sudden jumps or abrupt connections, thus generating a smooth, continuous motion path segment.

[0068] Step S1335: Sequentially concatenate the continuous action path segments to generate the guiding action path.

[0069] The continuous motion path segments obtained by fitting Bézier curves are spliced ​​together in their original order. During the splicing process, it is ensured that the start and end points of each continuous motion path segment accurately align with adjacent segments. Through this sequential splicing method, all continuous motion path segments are combined to form a complete guided motion path. This guided motion path includes the limb target coordinates and movement speed constraints at each time point, providing users with a motion guidance scheme with continuous movement logic. For example, continuous motion path segments P1, P2, P3, etc., are spliced ​​sequentially, so that the end point of P1 aligns with the start point of P2, and the end point of P2 aligns with the start point of P3, ultimately forming a complete guided motion path. This guided motion path can guide users to perform movements in a reasonable order and manner to achieve better therapeutic effects.

[0070] Step S134: For each time node of the guided action path, match the thermal image feedback area that overlaps with the current limb target coordinate space in the infrared thermal image, extract the temperature change of the corresponding thermal image feedback area when the action is executed in the historical interaction action trajectory data, and calculate the historical average value of the temperature change as the expected value of temperature change.

[0071] After generating the guided motion path, it is necessary to match the corresponding thermal feedback region for each time point and determine the expected temperature change value of that region. For each time point of the guided motion path, the thermal feedback region in the infrared thermal image that overlaps with the current limb target coordinate space is found. This requires spatially comparing the limb target coordinates in the guided motion path with each region in the infrared thermal image to identify the overlapping part as the thermal feedback region. For example, at a certain time point t, if the limb target coordinates in the guided motion path cover region R in the infrared thermal image, then region R is taken as the thermal feedback region for that time point.

[0072] Extract the temperature change of the thermal feedback area during the execution of corresponding actions from historical interaction action trajectory data. The historical interaction action trajectory data records the temperature changes of the thermal feedback area when the user performed various actions during past physical therapy training. For the thermal feedback area corresponding to each time point, find the temperature change of that area when performing the same or similar actions from historical data. For example, in historical data, when performing an action similar to the action at a certain moment in the guided action path, the temperature changes of the thermal feedback area R are ΔT1, ΔT2, ΔT3, etc. The historical average of these temperature changes is calculated as the expected temperature change value. The expected temperature change value is obtained by summing all relevant temperature changes and then dividing by the number of temperature changes. This expected temperature change value represents the predicted temperature change of the thermal feedback area when performing the corresponding action.

[0073] Step S135: Map the guided motion path, thermal feedback area, and expected temperature change value in three-dimensional space to generate a three-dimensional physiotherapy guidance model.

[0074] After obtaining the guided movement path, thermal feedback area, and expected temperature change value, these are mapped in three-dimensional space to generate a three-dimensional physiotherapy guidance model. Three-dimensional spatial mapping integrates and correlates this data in three-dimensional space, allowing it to be displayed intuitively. First, the coordinates of the limb targets in the guided movement path are located in three-dimensional space, forming the three-dimensional trajectory of the movement. Then, the thermal feedback area is mapped to the guided movement path in three-dimensional space, determining the three-dimensional position of the thermal feedback area corresponding to each movement node. Finally, the expected temperature change value is associated with the corresponding thermal feedback area and movement node, labeling each thermal feedback area with the expected temperature change value in three-dimensional space. Through this three-dimensional spatial mapping, the guided movement path, thermal feedback area, and expected temperature change value are integrated to generate a complete three-dimensional physiotherapy guidance model. This three-dimensional physiotherapy guidance model can intuitively display to the user the movement path to be performed during physiotherapy, the corresponding thermal feedback area, and the expected temperature change of that area, providing the user with a visualized physiotherapy guidance plan.

[0075] Step S140: Present a dynamic guidance interface of a three-dimensional physiotherapy guidance model to the user through a VR display device, simultaneously collect real-time interactive motion trajectory data of the user, and adjust the intensity of the motion prompts on the dynamic guidance interface based on the deviation between the real-time interactive motion trajectory data and the guidance motion path.

[0076] After generating the 3D physiotherapy guidance model, it needs to be presented to the user and dynamically adjusted based on the user's real-time movements. A dynamic guidance interface of the 3D physiotherapy guidance model is presented to the user through a VR display device. This interface can immersively display information such as the guided movement path and thermal feedback area, allowing the user to more intuitively understand the requirements of the physiotherapy movements. Simultaneously, real-time interactive movement trajectory data of the user is collected to monitor the user's actual movements.

[0077] Step S141: Render the dynamic guidance interface of the 3D physical therapy guidance model in the VR display device. The dynamic guidance interface includes a semi-transparent outline of the guidance action path and color indicators of the thermal image feedback area that are updated in real time.

[0078] A 3D physiotherapy guidance model is rendered in a VR display device to generate a dynamic guidance interface. This dynamic interface includes a semi-transparent outline of the guided movement path. The semi-transparency allows users to see both the guided movement path and their actual movement position, facilitating comparison and adjustment. Simultaneously, the interface includes real-time updated color indicators for the thermal imaging feedback area. The color of the thermal imaging feedback area updates in real-time according to temperature changes, allowing users to intuitively understand the temperature status of the area. For example, when the temperature of the thermal imaging feedback area rises, its color may change from blue to red to reflect the temperature change. Through this dynamic interface display, users can more clearly understand the physiotherapy requirements and real-time progress.

[0079] Step S142: Activate the motion capture module of the VR device to collect spatial coordinate data of the user's limb joints in real time and generate real-time interactive motion trajectory data.

[0080] The VR device's motion capture module is activated, enabling real-time monitoring of the spatial coordinates of the user's limb joints. By continuously tracking and recording the positions of these joints, real-time interactive motion trajectory data is generated. For example, the VR device can capture the spatial coordinates of joints such as the shoulder, elbow, and wrist using built-in sensors and cameras. As the user's movements change, the coordinate data of these joints is continuously updated, forming a continuous sequence of real-time interactive motion trajectory data. This real-time interactive motion trajectory data reflects the user's movements during actual physical therapy.

[0081] Step S143: Calculate the Euclidean distance between the limb coordinates at the current time node in the real-time interactive motion trajectory data and the target coordinates at the corresponding time node in the guided motion path, and use it as the deviation value.

[0082] After acquiring the real-time interactive motion trajectory data, it is necessary to calculate the deviation between it and the guided motion path. Specifically, this involves calculating the Euclidean distance between the limb coordinates at the current time point in the real-time interactive motion trajectory data and the target coordinates at the corresponding time point in the guided motion path.

[0083] Step S1431: Extract the set of limb joint coordinates for the current time node from the real-time interactive motion trajectory data. The set of limb joint coordinates includes the spatial coordinates of the shoulder, elbow, and wrist.

[0084] First, the set of limb joint coordinates for the current time point is extracted from the real-time interactive motion trajectory data. This set of limb joint coordinates includes the spatial coordinates of important joints such as the shoulder, elbow, and wrist. For example, at the current time point t, the real-time interactive motion trajectory data records the coordinates S_t of the shoulder joint, E_t of the elbow joint, and W_t of the wrist joint. These coordinates are combined to form the set of limb joint coordinates. This set of limb joint coordinates reflects the actual position of the user's limbs at the current moment.

[0085] Step S1432: Extract the target joint coordinate set corresponding to the time node in the guided motion path. The target joint coordinate set contains the spatial coordinates of the same parts as the limb joint coordinate set.

[0086] Next, the target joint coordinate set corresponding to the time node is extracted from the guided motion path. This target joint coordinate set contains the spatial coordinates of the same parts as the limb joint coordinate set in the real-time interactive motion trajectory data, namely the target coordinates of joints such as the shoulder, elbow, and wrist. For example, at the moment corresponding to the current time node t in the real-time interactive motion trajectory data, the guided motion path records the coordinates S_g of the shoulder target joint, E_g of the elbow target joint, and W_g of the wrist target joint. These coordinates are combined to form the target joint coordinate set. This target joint coordinate set represents the limb position that the user should reach at that moment.

[0087] Step S1433: Calculate the Euclidean distance of the spatial coordinates of each corresponding joint point to generate a set of joint point deviation values.

[0088] For each corresponding joint in the set of limb joint coordinates and the set of target joint coordinates, calculate the Euclidean distance between their spatial coordinates. For the shoulder joint, calculate the Euclidean distance D_S between the real-time shoulder coordinate S_t and the target shoulder coordinate S_g; for the elbow joint, calculate the Euclidean distance D_E between the real-time elbow coordinate E_t and the target elbow coordinate E_g; for the wrist joint, calculate the Euclidean distance D_W between the real-time wrist coordinate W_t and the target wrist coordinate W_g. Combine these Euclidean distances to form a set of joint deviation values. This set of joint deviation values ​​reflects the deviation of each user joint from the target position at the current moment.

[0089] Step S1434: Perform a weighted summation on the set of joint deviation values. The weighting coefficients are set according to the degree of influence of each joint on the physiotherapy effect to obtain the deviation values.

[0090] A weighted summation of the joint point deviation values ​​yields a comprehensive deviation value. The weighting coefficients are set based on the degree of influence of each joint point on the therapeutic effect; different joint points may have different importance during the therapeutic process. For example, the movement of the shoulder joint may have a greater impact on a particular therapeutic effect, so its deviation value will have a relatively large weighting coefficient; while the movement of the wrist joint has a relatively smaller impact on the therapeutic effect, so its deviation value will have a relatively small weighting coefficient. Each deviation value in the joint point deviation value set is multiplied by its corresponding weighting coefficient, and these products are then summed to obtain the final deviation value. This deviation value reflects the degree of deviation between the user's overall movement at the current moment and the guided movement path.

[0091] Step S144: Based on the comparison result between the deviation value and the preset deviation threshold, adjust the transparency and color saturation of the guide action path outline in the dynamic guide interface. The larger the deviation value, the lower the transparency and the higher the color saturation.

[0092] After obtaining the deviation value, it is compared with a preset deviation threshold. The preset deviation threshold is set based on actual physiotherapy requirements and movement accuracy standards. Based on the comparison results, the transparency and color saturation of the guided movement path outline in the dynamic guidance interface are adjusted. A larger deviation value indicates a greater deviation between the user's movement and the guided movement path. In this case, the transparency of the guided movement path outline is reduced to make it more noticeable, while the color saturation is increased to make the colors more vibrant. For example, when the deviation value exceeds a set percentage of the preset deviation threshold, the transparency of the guided movement path outline is reduced and the color saturation is increased. This more strongly prompts the user that their movement has a significant deviation and needs adjustment. Conversely, when the deviation value is small, the transparency of the guided movement path outline is high and the color saturation is low, avoiding excessive visual interference for the user.

[0093] Step S145: When the deviation value exceeds the third preset threshold, a limb adjustment direction arrow prompt is superimposed on the dynamic guidance interface, and the direction of the adjustment direction arrow points from the target coordinate to the opposite direction of the real-time limb coordinate.

[0094] If the deviation exceeds the third preset threshold, it indicates that the user's movement deviation has reached a relatively serious level, requiring more explicit prompts. In this case, a directional arrow indicating the adjustment direction is overlaid on the dynamic guidance interface. The arrow points from the target coordinates to the opposite direction of the real-time limb coordinates, visually indicating the direction the user should adjust their movement. For example, if the user's shoulder position deviates significantly from the target position in the guided movement path, and the deviation exceeds the third preset threshold, an arrow pointing from the target shoulder coordinates to the opposite direction of the real-time shoulder coordinates will be displayed on the dynamic guidance interface, prompting the user to adjust their shoulder in the direction indicated by the arrow to reduce the deviation from the guided movement path and improve the accuracy of the therapeutic movements.

[0095] Step S150: Based on the matching degree between the real-time interactive motion trajectory data and the guided motion path, and the difference between the actual temperature change value and the expected temperature change value of the thermal imaging feedback area, generate training effect feedback information that includes motion adjustment suggestions and thermal imaging optimization targets.

[0096] After a user completes a round of physical therapy training, the training effect needs to be evaluated and feedback information generated. This is achieved by analyzing the matching degree between real-time interactive motion trajectory data and guided motion paths, as well as the difference between the actual and expected temperature changes in the thermal imaging feedback area.

[0097] For example, step S150 includes the following steps: Step S151: Perform dynamic time warping matching of the real-time interactive action trajectory data and the guided action path over the entire time series, and calculate the path similarity between the real-time interactive action trajectory data and the guided action path as the matching degree.

[0098] Dynamic time warping matching is used to perform full-time series matching between real-time interactive action trajectory data and guided action paths. Dynamic time warping matching is a matching method capable of handling temporal scaling and deformation between time-series data. This method aligns and compares the real-time interactive action trajectory data and guided action paths along the time dimension to find the optimal matching path. The path similarity between the real-time interactive action trajectory data and the guided action path is calculated; this path similarity reflects the degree of similarity between the two in action patterns and temporal order, and is used as the matching degree. For example, by using the dynamic time warping matching algorithm, the path similarity between the real-time interactive action trajectory data and the guided action path is calculated to a certain value, which represents the degree of matching between the user's actual action and the guided action path.

[0099] Step S152: At the end of the training, re-collect the user's body surface temperature distribution data and extract the temperature difference of the thermal imaging feedback area before and after the training as the actual temperature change value.

[0100] At the end of training, the infrared sensor module is used again to collect data on the user's body surface temperature distribution. This newly collected data is compared with the data collected before training began, and the temperature difference in the thermal imaging feedback area before and after training is extracted as the actual temperature change value. For each thermal imaging feedback area, the difference between its temperature at the end of training and its temperature at the beginning of training is calculated to obtain the actual temperature change value for that area. For example, if the temperature of thermal imaging feedback area F at the beginning of training is T1, and the temperature at the end of training is T2, then the actual temperature change value for that area is ΔT = T2 - T1. This actual temperature change value reflects the actual temperature change of the thermal imaging feedback area during the training process.

[0101] Step S153: Calculate the absolute difference between the actual temperature change and the expected temperature change as the difference value.

[0102] After obtaining the actual temperature change value, it is compared with the previously determined expected temperature change value. The absolute difference between the actual and expected temperature change values ​​is calculated, and this absolute difference is taken as the discrepancy value. For example, if the actual temperature change value of the thermal imaging feedback area F is ΔT_actual, and the expected temperature change value is ΔT_expected, then the discrepancy value D = |ΔT_actual - ΔT_expected|. This discrepancy value reflects the degree of deviation between the actual and expected temperature changes in the thermal imaging feedback area.

[0103] Step S154: Based on the correspondence between the matching degree and the preset matching degree level, generate motion adjustment suggestions. The motion adjustment suggestions include the time nodes of the motions that need to be corrected and the spatial coordinate range of the limbs.

[0104] Based on the correspondence between the matching degree and the preset matching degree level, movement adjustment suggestions are generated. The preset matching degree level is divided into different levels according to actual physical therapy requirements and movement accuracy standards. For each matching degree level, there are corresponding movement adjustment suggestion rules. For example, when the matching degree is at a low level, it indicates that the user's movement deviates significantly from the guided movement path, requiring more comprehensive movement adjustments. By analyzing the matching between real-time interactive movement trajectory data and the guided movement path, the time nodes and limb spatial coordinate ranges that need to be focused on correction are identified. For example, at certain time nodes, the user's shoulder, elbow, and other joint movements deviate significantly. The movement adjustment suggestions will clearly indicate these time nodes and the limb spatial coordinate ranges that need adjustment, guiding the user to improve their movements.

[0105] Step S155: Based on the correspondence between the difference value and the preset difference value range, generate a thermal image optimization target. The thermal image optimization target includes the amount of temperature change that needs to be increased and the corresponding thermal image feedback area location.

[0106] After obtaining the difference between the actual and expected temperature changes, the difference is mapped to a preset difference range. This preset difference range is divided into different intervals based on the expected therapeutic effect and evaluation criteria, with each interval corresponding to a different thermal imaging optimization goal. For each difference value, its corresponding preset difference range is determined, and then a thermal imaging optimization goal is generated according to the rules for that range. For example, when the difference value is in a large range, it indicates that the actual temperature change in the thermal imaging feedback area differs significantly from the expected temperature change, requiring a substantial increase in the temperature change. By analyzing the location and characteristics of the thermal imaging feedback area, the required increase in temperature change and the corresponding location of the thermal imaging feedback area are determined. Specifically, it can be clearly identified which thermal imaging feedback areas need temperature increases and the approximate range of temperature change increases required. For instance, if the temperature change difference between thermal imaging feedback areas G and H is large, the thermal imaging optimization goal will specify that a set range of temperature change needs to be increased in thermal imaging feedback area G, and will also specify the required range of temperature change increases for thermal imaging feedback area H, as well as the specific location information of these two areas, for targeted adjustments to the therapeutic treatment later.

[0107] Step S156: Align the motion adjustment suggestions and thermal imaging optimization targets with timestamps to generate training effect feedback information.

[0108] After generating motion adjustment suggestions and thermal imaging optimization goals separately, they need to be timestamped. Since motion adjustment suggestions and thermal imaging optimization goals may involve different time points, their temporal consistency needs to be ensured to make the feedback information clearer and more coherent. Specifically, using the time points in the training process as a benchmark, information from the same time points in the motion adjustment suggestions and thermal imaging optimization goals is associated and integrated. For example, at a certain time point t, the motion adjustment suggestion indicates the need to adjust shoulder movements, while the thermal imaging optimization goal indicates that the corresponding thermal feedback area needs to increase the set temperature change amount. These two pieces of information are aligned and stitched together at time point t. Through this timestamping alignment method, the motion adjustment suggestions and thermal imaging optimization goals are combined to generate training effect feedback information that includes both. This training effect feedback information comprehensively reflects the user's performance in physical therapy training and areas for improvement, providing clear guidance to help the user improve training effects in subsequent physical therapy training and achieve better physical therapy goals.

[0109] Throughout the data collection and processing process, sensitive data such as user body surface temperature distribution and historical interaction trajectory data are involved. Multiple encryption technologies are employed to protect this data. During data collection, data acquired by the infrared sensing module and VR device is immediately encrypted using a symmetric encryption algorithm to ensure data security during transmission. In the data storage phase, the encrypted data is stored on a secure server with access control policies, allowing only authorized personnel to access the data. Regular backups are also performed to prevent data loss. During data processing, the data is anonymized, removing user-identifiable information and retaining only features relevant to physical therapy training, thus reducing the risk of data leakage. When using AI models for processing, the training and inference processes are conducted in a secure environment to prevent data leakage during model processing. These technical measures effectively protect the security of users' sensitive privacy data and avoid the risk of data leakage.

[0110] Figure 2 The illustration shows exemplary hardware and software components of a VR-interactive infrared physiotherapy visualization-guided training system 100, which can implement the ideas of this application, according to some embodiments of this application. For example, a processor 120 can be used in the VR-interactive infrared physiotherapy visualization-guided training system 100 and to perform the functions in this application.

[0111] The VR-interactive infrared physiotherapy visualization-guided training system 100 can be a general-purpose server or a special-purpose server; both can be used to implement the VR-interactive infrared physiotherapy visualization-guided training method of this application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0112] For example, the VR-interactive infrared physiotherapy visualization-guided training system 100 may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the VR-interactive infrared physiotherapy visualization-guided training system 100 may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The VR-interactive infrared physiotherapy visualization-guided training system 100 also includes an I / O interface 150 between the computer and other input / output devices.

[0113] For ease of explanation, only one processor is described in the VR-interactive infrared physiotherapy visualization-guided training system 100. However, it should be noted that the VR-interactive infrared physiotherapy visualization-guided training system 100 of this application may also include multiple processors. Therefore, the steps executed by one processor as described in this application may also be executed jointly or individually by multiple processors. For example, if the processor of the VR-interactive infrared physiotherapy visualization-guided training system 100 executes steps A and B, it should be understood that steps A and B may also be executed jointly by two different processors or individually by one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor jointly execute steps A and B.

[0114] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned VR-interactive infrared physiotherapy visualization-guided training method is implemented.

[0115] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A VR-interactive, visualized, guided training method for infrared physiotherapy, characterized in that, The method includes: The system acquires the user's current body surface temperature distribution data and the historical interaction trajectory data recorded by the VR device. The body surface temperature distribution data is acquired through an infrared sensing module, and the historical interaction trajectory data includes the user's limb spatial coordinate sequence at each time point in past physical therapy training. An infrared thermal image covering the target physiotherapy area is generated based on the body surface temperature distribution data. The infrared thermal image contains temperature gradient stratification information and low-temperature sensitive area markings based on a preset threshold. The historical interactive action trajectory data and the infrared thermal image are input into the physiotherapy model construction module. Through spatial location correlation analysis, a three-dimensional physiotherapy guidance model containing the guided action path and thermal image feedback area is generated. In the three-dimensional physiotherapy guidance model, each guided action node corresponds to the expected temperature change value of a thermal image feedback area. The dynamic guidance interface of the three-dimensional physiotherapy guidance model is presented to the user through a VR display device. Real-time interactive action trajectory data of the user is collected simultaneously. The intensity of the action prompts of the dynamic guidance interface is adjusted based on the deviation between the real-time interactive action trajectory data and the guidance action path. Based on the matching degree between the real-time interactive motion trajectory data and the guided motion path, and the difference between the actual temperature change value and the expected temperature change value of the thermal imaging feedback area, training effect feedback information containing motion adjustment suggestions and thermal imaging optimization goals is generated.

2. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 1, characterized in that, The acquisition of the user's current body surface temperature distribution data and the historical interaction trajectory data recorded by the VR device includes: The multi-array infrared sensor installed in the physiotherapy chamber is activated to continuously scan the temperature of the user's target physiotherapy area, obtaining multiple frames of raw temperature data with timestamps. The raw temperature data is calibrated by spatial coordinates, and the sensor coordinates of each frame of raw temperature data are mapped to the user's body surface anatomical coordinate system to generate the body surface temperature distribution data. Retrieve the user's historical physiotherapy training logs stored in the VR device, and extract the action recording segments that contain the spatial coordinates of limb joints. The time span of the action recording segments is the most recent preset training cycle. The action recording segments are time-aligned to generate historical interaction action trajectory data containing a continuous time series. The temporal resolution of the historical interaction action trajectory data is consistent with the scanning frequency of the body surface temperature distribution data.

3. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 2, characterized in that, The step of performing spatial coordinate calibration on the raw temperature data, mapping the sensor coordinates of each frame of raw temperature data to the user's anatomical coordinate system, includes: Acquire three-dimensional scan data of the user's body surface anatomical structure and construct a body surface coordinate system containing key anatomical landmarks; Extract temperature sampling points from the original temperature data that correspond to the spatial locations of the key anatomical landmarks; Calculate the coordinate transformation matrix between the sensor coordinates of the temperature sampling point and the coordinates of the corresponding anatomical landmark in the body surface coordinate system. The coordinate transformation matrix includes a translation vector and a rotation matrix. The coordinate transformation matrix is ​​used to transform the coordinates of all sampling points of the original temperature data to generate the body surface temperature distribution data mapped to the body surface anatomical coordinate system.

4. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 1, characterized in that, The step of generating an infrared thermal image covering the target physiotherapy area based on the body surface temperature distribution data, wherein the infrared thermal image contains temperature gradient stratification information and low-temperature sensitive area markers based on a preset threshold, includes: The surface temperature distribution data is interpolated and filled in with temperature values ​​in the missing data areas to generate continuous temperature field data that fully covers the target physiotherapy area. Calculate the rate of temperature change of adjacent spatial points based on the continuous temperature field data, and construct a temperature gradient field model based on the rate of temperature change. According to the preset temperature range division rules, the continuous temperature field data is mapped to the thermal image color coding table to generate temperature gradient layering information with layered color identification. Extract the boundaries of regions with temperature values ​​below a first preset threshold from the continuous temperature field data, connect the discrete low-temperature regions through morphological closing operations, and generate the low-temperature sensitive region markers. The low-temperature sensitive region markers include a region contour coordinate sequence and the average temperature value within the region.

5. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 4, characterized in that, The interpolation and completion process for the continuous temperature field data, filling in the temperature values ​​in the missing data regions, includes: Identify missing regions in the body surface temperature distribution data, where the missing regions are a set of continuous spatial points whose temperature values ​​are marked as invalid; Extract the effective temperature values ​​of the boundary of the missing region and construct a temperature prediction function based on bicubic spline interpolation; The spatial coordinates of the missing region are input into the temperature prediction function to calculate the predicted temperature value of the missing region. The predicted temperature values ​​are filled into the missing regions to generate the continuous temperature field data.

6. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 1, characterized in that, The step of inputting the historical interactive action trajectory data and the infrared thermal image into the physiotherapy model construction module, and generating a three-dimensional physiotherapy guidance model containing the guided action path and thermal feedback area through spatial location correlation analysis, includes: Principal component analysis was performed on the historical interaction trajectory data to extract the user's high-frequency repetitive action patterns as basic action units; The spatial overlap between the limb spatial coordinates of the basic action unit and the contour coordinates of the low-temperature sensitive area of ​​the infrared thermogram is calculated, and basic action units with an overlap of more than a second preset threshold with the low-temperature sensitive area are selected as candidate guided actions. The candidate guided actions are reassembled in time series to generate a guided action path with continuous motion logic. The guided action path includes the limb target coordinates and action speed constraints at each time node. For each time node of the guided action path, match the thermal image feedback area in the infrared thermal image that overlaps with the coordinate space of the current limb target, extract the temperature change of the thermal image feedback area when the corresponding action is executed from the historical interactive action trajectory data, and calculate the historical average value of the temperature change as the expected value of the temperature change. The guided action path, thermal imaging feedback area, and expected temperature change value are mapped in three-dimensional space to generate the three-dimensional physiotherapy guidance model.

7. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 6, characterized in that, The step of reorganizing the candidate guiding actions into a time series to generate a guiding action path with continuous motion logic includes: Extract the start and end coordinates of the candidate guided actions to construct motion direction vectors between action nodes; Calculate the spatial distance between the termination coordinates of adjacent candidate guide actions and the starting coordinates of the next candidate guide action, and filter candidate guide action pairs whose spatial distance is less than a fourth preset threshold. The time interval of the candidate guiding action pair is adjusted so that the end time of the previous action is continuously connected with the start time of the next action. By fitting the motion trajectory of the candidate guided action pair with a Bézier curve, a smooth transition continuous motion path segment is generated. The continuous action path segments are sequentially spliced ​​together to generate the guided action path.

8. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 1, characterized in that, The process of presenting the three-dimensional physiotherapy guidance model to the user via a VR display device, simultaneously collecting real-time interactive motion trajectory data of the user, and adjusting the intensity of the motion prompts on the dynamic guidance interface based on the deviation between the real-time interactive motion trajectory data and the guidance motion path includes: A dynamic guidance interface for rendering the three-dimensional physiotherapy guidance model is provided in a VR display device. The dynamic guidance interface includes a semi-transparent outline of the guidance action path and color indicators of the thermal image feedback area that are updated in real time. The motion capture module of the VR device is activated to collect the spatial coordinate data of the user's limb joints in real time and generate the real-time interactive motion trajectory data. Calculate the Euclidean distance between the limb coordinates at the current time node in the real-time interactive motion trajectory data and the target coordinates at the corresponding time node in the guided motion path, and use it as the deviation value; Based on the comparison result between the deviation value and the preset deviation threshold, the transparency and color saturation of the guide action path outline in the dynamic guide interface are adjusted. The larger the deviation value, the lower the transparency and the higher the color saturation. When the deviation value exceeds the third preset threshold, a limb adjustment direction arrow prompt is superimposed on the dynamic guidance interface. The direction of the adjustment direction arrow is from the target coordinate to the opposite direction of the real-time limb coordinate.

9. The VR-interactive infrared physiotherapy visualization-guided training method according to claim 8, characterized in that, The calculation of the Euclidean distance between the limb coordinates at the current time point in the real-time interactive motion trajectory data and the target coordinates at the corresponding time point in the guided motion path, as the deviation value, includes: Extract the set of limb joint coordinates for the current time node from the real-time interactive motion trajectory data. The set of limb joint coordinates includes the spatial coordinates of the shoulder, elbow, and wrist. Extract the target joint coordinate set corresponding to the time node in the guided action path. The target joint coordinate set includes the spatial coordinates of the same part as the limb joint coordinate set. Calculate the Euclidean distance of the spatial coordinates of each corresponding joint point to generate a set of joint point deviation values; The set of joint deviation values ​​is weighted and summed, with the weighting coefficients set according to the degree of influence of each joint on the physiotherapy effect, to obtain the deviation values.

10. A VR-interactive infrared physiotherapy visualization-guided training system, characterized in that, The device includes a processor and a memory, the memory being connected to the processor. The memory is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the memory to implement the VR-interactive infrared physiotherapy visualization-guided training method as described in any one of claims 1-9.

Images