Intelligent skin wound area measuring and analyzing terminal

Abstract

The invention discloses an intelligent skin wound area measurement and analysis terminal, and relates to the technical field of medical instruments. The invention aims to solve the problems that the traditional flexible ruler easily causes cross infection, has large error on irregular and sensitive wounds, cannot capture three-dimensional characteristics such as wound depth to judge severity and healing potential, lacks healing risk pre-judgment ability and often misses good intervention opportunities. According to the invention, the laser projection part, the non-contact depth sensor and the optical imaging camera cooperatively collect data, so that cross infection caused by direct contact of a traditional flexible rule with a wound is avoided; meanwhile, the area is accurately calculated through a grid scale and three-dimensional point cloud data, a complete wound three-dimensional model is generated through a wound model building module, three-dimensional characteristic parameters such as depth, texture and color proportion are quantitatively extracted, a multi-dimensional objective basis is provided for judging the severity and healing potential of the wound, future healing trend prediction and risk early warning are generated, and the wound healing effect is improved. And medical workers are helped to avoid risks in advance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical devices, in particular to a skin wound area intelligent measurement and analysis terminal. BACKGROUND

[0002] In clinical nursing, accurate assessment of skin wounds is the core basis for developing treatment plans, tracking healing progress and evaluating efficacy. However, traditional soft ruler measurement relies on direct operation of medical staff on the wound surface, which not only easily causes cross-infection of the wound and measurement tools, but also is difficult to obtain accurate length x width data for irregular shapes or sensitive wounds, with large measurement errors. Although the existing projection ruler realizes non-contact area measurement, it can only provide two-dimensional plane data and cannot capture three-dimensional features such as wound depth, concave level and tissue defect, so it cannot judge wound severity and healing potential according to these indicators. It lacks the ability to predict wound healing risks, and often adjusts the plan only after problems such as infection and delayed healing appear, missing the opportunity for intervention. SUMMARY

[0003] The purpose of the present application is to provide a skin wound area intelligent measurement and analysis terminal, which synchronously collects data through a laser projection component, a non-contact depth sensor and an optical imaging camera, without relying on regular wound morphology, accurately calculates the area through a grid ruler and three-dimensional point cloud data, generates future healing trend prediction and risk warning, helps medical staff to avoid risks in advance and improves treatment efficiency, to solve the problems raised in the above background technology.

[0004] To achieve the above purpose, the present application provides the following technical scheme: The skin wound area intelligent measurement and analysis terminal comprises a handheld intelligent measurement terminal, wherein the handheld intelligent measurement terminal is internally integrated with a wound data acquisition module, a wound model construction module and a wound analysis module. The wound data acquisition module is used for synchronously capturing wound data information based on a depth sensor and an optical imaging camera, processing and integrating the acquired wound data information, and generating corresponding three-dimensional point cloud data. The wound model construction module is used for preprocessing the three-dimensional point cloud data, constructing a complete wound three-dimensional model based on the corresponding three-dimensional point cloud data of multiple perspectives, and determining a wound edge contour. The wound analysis module is used for calling historical wound measurement data of the same patient, generating a dynamic healing curve based on a time sequence, labeling healing key nodes, integrating multi-dimensional information, and generating a wound later change prediction report and adjustment suggestion.

[0005] Further, the wound data acquisition module comprises: The collection calibration unit is configured to calibrate each component before collecting skin wound data, and determine a laser projection mode and collection parameters according to pre-input operation requirements and wound type information after calibration is completed. The data synchronization capturing unit is configured to synchronously acquire wound area, original depth data and original optical data of a patient wound area based on the determined laser projection mode and collection parameters, and monitor working synchronization of each component in real time during the collection process. The wound data integration unit is configured to pre-process the acquired original depth data and original optical data, establish a corresponding relationship between the original optical data and the original depth data, extract wound color feature data in combination with the projection scale data, and generate three-dimensional point cloud data.

[0006] Further, the wound data collection module further comprises: The depth sensor is configured to extract a collection reference distance between the handheld intelligent measurement terminal and the surface of the skin wound area, and simultaneously extract an imaging distance of the optical imaging camera, and compare and calibrate the imaging distance and the collection reference distance. The calibrated collection reference distance is matched with a preset laser projection adaptation threshold value to calculate a matching degree, and whether the current distance is in an effective measurement interval is determined based on a calculation result. An adaptation relationship between the collection reference distance and the grid density is established, an adaptation distance parameter of each component is determined in combination with a preset adaptation distance parameter reference value, and the collection parameters of each component are set based on the adaptation distance parameter.

[0007] Further, the data synchronization capturing unit further comprises: The grid-shaped projection scale and the corresponding skin damage color contrast chart are transmitted on the patient wound and the surrounding normal skin area according to the determined laser projection mode, and the corresponding components are controlled to scan and collect the patient wound area according to the preset collection parameters. The laser projection component continuously projects the grid-shaped projection scale, counts the number of complete grids and half grids covering the wound, acquires preliminary area data of the wound in real time, compares the actual color of the wound with the projected standard color contrast chart, and determines the color type of the wound. The depth sensor acquires depth information of different positions of the wound in real time, and the optical imaging camera synchronously acquires image data of the patient wound area, and synchronously acquires original depth data of internal concave levels and tissue defects of the wound and original optical data containing surface texture, edge contour, projection scale and color contrast chart information of the wound.

[0008] Further, the wound model construction module comprises: The edge contour determination unit is configured to extract a depth difference, a color difference and a texture difference between the wound area and the surrounding normal skin area in the pre-processed three-dimensional point cloud data, and construct a boundary judgment index of multi-feature fusion. The preset normal skin pixel points are taken as seed points, and the normal skin region is gradually grown and expanded in combination with the boundary judgment index to determine the growth boundary of the region not covered by the growth.

[0009] The growth boundary is smoothed to generate continuous wound edge contour data, and the area of the region surrounded by the contour is calculated. The wound area calculated by the wound data acquisition module is subtracted, and if the difference exceeds a preset area deviation threshold, the three-dimensional point cloud data preprocessing operation is performed again. The model output unit is configured to construct a complete three-dimensional wound model based on the determined wound edge contour, in combination with the original optical data and the original depth data, and to synchronously transmit the constructed complete three-dimensional wound model and the wound edge contour data to the wound analysis module.

[0010] Further, the wound model construction module further includes a wound feature quantization unit configured to perform feature quantization extraction on the complete three-dimensional wound model to generate multi-dimensional wound feature parameters. Based on the edge contour generated by the wound edge contour determination unit, the actual wound area is obtained based on the coordinates of the discrete points of the contour. Each surface point in the three-dimensional wound model is traversed, the perpendicular distance of each surface point from the normal skin surface around the wound is calculated, the maximum distance is recorded as the maximum depth of the wound, and the average of the distances of all the concave points is calculated as the average depth of the concave points. The gray level co-occurrence matrix of the wound surface model is extracted, and the texture feature parameters such as contrast, correlation, and energy of the gray level co-occurrence matrix are calculated to quantitatively represent the roughness and uniformity of the wound surface. Based on the wound color feature data, the wound area is divided into multiple sub-areas, the color features of each sub-area are extracted and counted, the proportion of each color type in the wound area is obtained, and a color distribution histogram is generated.

[0011] Further, the wound analysis module includes: The historical data retrieval unit is configured to receive patient information input by medical staff, retrieve historical wound measurement data of the patient, including historical three-dimensional wound models, multi-dimensional wound feature parameters, and corresponding measurement time stamps, and construct a wound data time series corresponding to the patient. The healing curve generation unit is configured to determine target values of the multi-dimensional wound feature parameters based on the wound data time series, and to draw an area healing curve, a depth healing curve, and a color healing curve of the patient based on the target values. The healing prediction and early warning unit is configured to extract multi-dimensional information features of the current patient, perform feature matching on standardized cases in a clinical healing case database, filter reference cases, and generate a corresponding prediction report for risk early warning, while generating corresponding intervention suggestions. The report output unit is configured to integrate the dynamic healing curve, the future healing trend prediction curve, the risk early warning result and the intervention suggestion into a wound late-stage change prediction report.

[0012] Further, the early warning step of the risk early warning in the healing prediction early warning unit comprises: According to the corresponding relationship between the target value and the historical measurement points in the historical wound measurement data, the historical healing rates of each healing type are calculated; According to the historical healing rates, the historical healing rate mean and the historical healing rate standard deviation of each healing type are calculated; According to the historical healing rate mean and the historical healing rate standard deviation, the dynamic risk threshold of each healing type is calculated; If the current healing rate is less than the corresponding dynamic risk threshold, a delayed healing warning of the corresponding healing type is generated, and the delayed healing time is calculated; The delayed healing warning and the corresponding delayed healing time are jointly used as risk information for warning.

[0013] Further, the handheld intelligent measurement terminal body is further integrated with a light supplement module and a power module, wherein the light supplement module can automatically adjust the brightness according to the ambient light intensity, and the power module is a rechargeable lithium battery.

[0014] Further, the light supplement module further comprises: According to the collection reference distance of the handheld intelligent measurement terminal and the surface of the skin wound area, a terminal three-dimensional model is constructed in the three-dimensional space of the wound three-dimensional model; According to the light supplement strategy preset by the light supplement device, light supplement is simulated in the three-dimensional space, and the first direction vector from the light source center of the virtual light supplement device to each surface point under each light supplement strategy is determined; According to the perpendicular line from the surface point to the normal skin surface around the wound, a second direction vector is constructed; The vector angle of the first direction vector and the second direction vector is calculated and used as the light supplement angle of the corresponding surface point; According to the preset light supplement angle suitability evaluation library, the first suitable value of the light supplement angle is matched and associated with the surface point; The illumination parameter of each surface point under each light supplement strategy is determined; Based on the light supplement suitability evaluation model, the second suitable value of the illumination parameter is determined according to the color features and texture features of the surface point and is associated with the surface point; the light supplement suitability evaluation model is obtained by training a plurality of artificial evaluation records that evaluate whether the known illumination parameter is conducive to the clear presentation of the wound features according to the wound images labeled with color features and texture features; during training, the labeled color features, texture features and known illumination parameter are used as input parameters of the machine learning model, and the artificial evaluation value is used as the output parameter; The first suitable value and the second suitable value of the surface point are fused by weighting to obtain a target suitable value corresponding to the surface point; The target suitable values corresponding to all surface points under the same light supplement strategy are summed to obtain a suitable degree corresponding to the light supplement strategy; The light supplement strategy with the highest suitable degree is applied.

[0015] Compared with the prior art, the beneficial effects of the present application are: The laser projection component cooperates with the non-contact depth sensor and the optical imaging camera to collect data, avoiding cross infection caused by direct contact of the traditional soft ruler with the wound; at the same time, it does not need to rely on the regular shape of the wound, and even for irregular and sensitive wounds, the grid ruler and three-dimensional point cloud data can accurately calculate the area, solve the problem of large traditional measurement error, generate a complete three-dimensional wound model through the wound model construction module, quantitatively extract stereoscopic feature parameters such as depth, texture and color ratio, provide multi-dimensional objective basis for judging the severity and healing potential of the wound, generate future healing trend prediction and risk warning, and output intervention suggestions to help medical staff avoid risks in advance. BRIEF DESCRIPTION OF DRAWINGS

[0016] Fig. 1 The figure is a skin wound area intelligent measurement and analysis terminal module diagram of the present application; Fig. 2 The figure is a wound data acquisition module projection scanning schematic diagram of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] In order to solve the technical problems that the traditional soft ruler measurement is easy to cause cross infection and has large error for irregular and sensitive wounds, the existing projection ruler can only provide two-dimensional data for non-contact area measurement, cannot capture the depth and other stereoscopic features of the wound to judge the severity and healing potential, and lacks healing risk prediction ability, often missing the opportunity to intervene, please refer to Figs. 1-2 The present embodiment provides the following technical solutions: The skin wound area intelligent measurement and analysis terminal comprises a handheld intelligent measurement terminal, wherein the handheld intelligent measurement terminal is internally integrated with a wound data acquisition module, a wound model construction module and a wound analysis module; an arc shape is adopted at the edge of the handheld intelligent measurement terminal to adapt to the palm holding arc and ensure the portability of medical staff during mobile ward rounds and bedside operations; the handheld intelligent measurement terminal body is further internally integrated with a light supplement module and a power module, wherein the light supplement module is an adjustable color temperature LED arranged on both sides of the data acquisition module, the color temperature range is 3000K-6500K, the brightness can be automatically adjusted according to the environmental light intensity, and the data acquisition module can obtain clear wound image and depth information under different illumination scenes such as natural light in the ward and surgical shadowless lamp, and the power module is a rechargeable lithium battery, which meets the power demand of medical staff during all-day mobile diagnosis and treatment; The wound data acquisition module is internally provided with a high-resolution depth sensor, an optical imaging camera and laser projection components on both sides of the camera, is used for synchronously capturing wound data information including surface texture, edge contour, internal recess level and tissue defect information based on the depth sensor and the optical imaging camera, processing and integrating the obtained wound data information, and generating corresponding three-dimensional point cloud data, wherein the high-resolution depth sensor is integrated in the top region of the handheld intelligent measurement terminal body, the optical imaging camera synchronously collects high-definition wound surface images, and the spatial position matching of the three-dimensional point cloud data and the optical images is ensured: The wound model construction module is used for pre-processing the three-dimensional point cloud data, sequentially performing noise reduction, splicing and reconstruction processing, constructing a complete wound three-dimensional model based on the three-dimensional point cloud data corresponding to multiple views, and determining a wound edge contour; The wound analysis module is used for calling historical wound measurement data of the same patient, generating a dynamic healing curve based on a time sequence, labeling healing key nodes, integrating multi-dimensional information, generating a wound late change prediction report and adjustment suggestion, and the like; In the embodiment, the multi-dimensional information comprises historical measurement data, wound type, patient basic health condition and treatment scheme; the wound type comprises pressure injury, diabetic foot ulcer, burn wound and the like; the patient basic health condition comprises age, blood sugar level and nutrition state; and the treatment scheme comprises dressing type and medication condition; In the embodiment, the clinical healing case database contains standardized wound healing cases, and a feature matching is performed through a random forest, a neural network and the like machine learning algorithm to generate a wound healing trend prediction and risk early warning for future 7-14 days, such as healing delay probability and infection risk level.

[0019] In the embodiment, the non-contact acquisition mode is adopted to avoid the cross-infection risk caused by the direct contact of the traditional soft ruler, and the three-dimensional point cloud data is used to capture the three-dimensional features such as wound depth and concave level, so as to solve the limitation of the existing projection ruler that can only provide two-dimensional data, accurately judge the wound severity and healing potential, automatically generate a dynamic healing curve and key node label, so that medical staff can quickly judge the treatment effect without manually arranging data, and the multi-dimensional information is integrated to generate a dynamic curve and adjustment suggestion, which provides a scientific basis for treatment plan making, healing trend prediction and risk warning, so that medical staff have sufficient time to adjust the plan.

[0020] In the embodiment, the wound data acquisition module comprises: The calibration unit is used to calibrate each component before collecting skin wound data, including a depth sensor, an optical imaging camera and a laser projection component; first, the spatial coordinate matching calibration of the high-resolution depth sensor and the optical imaging camera is performed to eliminate the initial deviation of the two; then the laser projection component is controlled to project a grid-shaped projection ruler and a standard color contrast chart to the preset calibration reference, the calibration reference image after projection is collected by the optical imaging camera, the size deviation of the projection ruler and the color deviation of the color contrast chart are analyzed, and the projection angle, focal length, color temperature and brightness of the laser projection component are automatically adjusted until the deviation is within the preset precision threshold; after calibration, the laser projection mode and the collection parameters are determined according to the pre-input operation requirements and wound type information; The data synchronization capture unit is used to synchronize the collection of the wound area, the original depth data and the original optical data of the patient's wound area based on the determined laser projection mode and collection parameters, and to monitor the working synchronization of each component in real time during the collection process. The collection / projection time stamps of the three are compared, and if the time difference exceeds the preset synchronization threshold, the collection is immediately paused and the start timing is recalibrated. After calibration, the collection is continued to ensure that the original depth data and the original optical data containing projection information correspond to each other in the time dimension; In the embodiment, the data synchronization capture unit further comprises: According to the determined laser projection mode, the grid-shaped projection ruler and the corresponding skin damage color contrast chart are transmitted on the patient's wound and the surrounding normal skin area, and the corresponding components are controlled to scan and collect the patient's wound area according to the preset collection parameters; The laser projection component continuously projects the grid-shaped projection ruler, counts the number of complete grids and half grids covering the wound, obtains the preliminary area data of the wound in real time, compares the actual color of the wound with the projected standard color contrast chart, and determines the color type of the wound; The depth sensor collects depth information of different positions of the wound in real time, and the optical imaging camera synchronously collects image data of the wound area of the patient, so as to synchronously obtain original depth data of a concave level and tissue defect inside the wound and original optical data containing surface texture, edge contour, projection scale and color contrast map information of the wound surface; The wound data integration unit is configured to pre-process the obtained original depth data and original optical data, including removing outliers and interpolating and completing missing points for the original depth data, performing noise reduction processing and image segmentation for the original optical data, separating the wound area, the projection scale area and the color contrast map area, establishing a corresponding relationship between the original optical data and the original depth data, calculating the actual area of the wound in combination with the projection scale data, extracting the color feature data of the wound in combination with the color contrast map data, and fusing the optical information of the pixel points and the depth information of the depth data points to generate three-dimensional point cloud data containing the area, color feature, surface texture, edge contour, internal concave level and tissue defect information of the wound.

[0021] In the embodiment, the wound data acquisition module further includes: The depth sensor extracts a collection reference distance of the handheld intelligent measurement terminal and the surface of the skin wound area, and simultaneously extracts an imaging distance of the optical imaging camera, and the imaging distance is compared and calibrated with the collection reference distance. The calibrated collection reference distance is matched with a preset laser projection adaptation threshold to calculate a matching degree, and whether the current distance is in an effective measurement interval is determined based on a calculation result, so as to ensure that the distance range in which the projection scale completely covers the wound area and the grid is clear and distinguishable. An adaptation relationship between the collection reference distance and the grid density is established, an adaptation distance parameter of each component is determined in combination with a preset adaptation distance parameter reference value, and the collection parameters of each component are set based on the adaptation distance parameter. In the embodiment, the collection reference distance is an initial detection distance of the depth sensor and the surface of the skin wound area, and an initial value is preset based on a conventional measurement range of the skin wound. In the embodiment, the imaging distance is a straight-line distance between the center of the lens of the optical imaging camera and the center point of the skin wound area, and the imaging distance is consistent with the collection reference distance of the high-resolution depth sensor. In the embodiment, the sampling frequency of the high-resolution depth sensor is adjusted according to the wound concave depth detection result, the sampling frequency is higher when the wound concave is deeper, and the concave level data collection is complete; the imaging resolution of the optical imaging camera is adjusted according to the size of the wound area, the imaging resolution is higher when the wound area is smaller, and the surface texture of the wound is clear; and the projection focal length of the laser projection component is adjusted according to the adaptation distance parameter, so that the grid-shaped projection scale is clearly imaged in the wound area without edge distortion or blur, thereby providing parameter support for accurate collection of multi-dimensional data of the wound.

[0022] In the embodiment, by synchronously extracting distance data of the depth sensor and the optical camera, combining with the laser projection characteristics to calculate the adaptive distance parameter, the sensor combination can be ensured to maintain the optimal measurement distance with the wound, effectively avoiding the problems of insufficient collection range caused by too close distance or data precision decline caused by too far distance, realizing automatic synchronization of distance data and automatic calculation of adaptive parameters, without manual measurement and adjustment by medical staff, adapting to the scene requirements of handheld terminal mobile ward round and bedside operation, significantly improving the operation convenience and data accuracy, optimizing the depth sensor sampling frequency, camera resolution and laser projection focal length, accurately matching the collection parameters for different types of wounds, ensuring that the wound information in each dimension can be completely captured, and avoiding collection omission or distortion caused by fixed parameters.

[0023] In the embodiment, the wound model construction module comprises: An edge contour determination unit is configured to extract the depth difference, color difference and texture difference between the wound area and the surrounding normal skin area in the preprocessed three-dimensional point cloud data, and construct a boundary judgment index based on multi-feature fusion. The preset normal skin pixel points are taken as seed points, and the normal skin area is gradually grown and expanded based on the boundary judgment index, and the growth boundary of the area not covered by the growth is determined.

[0024] The growth boundary is smoothed to generate continuous wound edge contour data, and the area of the contour-enclosed region is calculated, and the difference between the wound area calculated by the wound data collection module is calculated, and if the difference exceeds the preset area deviation threshold, the three-dimensional point cloud data preprocessing operation is performed again. A model output unit is configured to construct a complete wound three-dimensional model based on the determined wound edge contour, combined with the original optical data and the original depth data, and synchronously transmit the constructed complete wound three-dimensional model and the wound edge contour data to the wound analysis module.

[0025] In the embodiment, the wound model construction module further comprises a wound feature quantization unit configured to perform feature quantization extraction on the complete wound three-dimensional model to generate multi-dimensional wound feature parameters. Based on the edge contour generated by the wound edge contour determination unit, the actual area of the wound is obtained based on the coordinates of the contour discrete points. Each surface point in the three-dimensional wound model is traversed, the perpendicular distance between each surface point and the surface of the normal skin around the wound is calculated, the maximum distance is recorded as the maximum depth of the wound, and the distance average of all recessed points is calculated and taken as the average depth of the recess. The gray level co-occurrence matrix of the wound surface model is extracted, and the texture feature parameters such as contrast, correlation and energy of the gray level co-occurrence matrix are calculated to quantitatively represent the roughness and tissue uniformity of the wound surface. Based on the wound color feature data, the wound area is divided into multiple sub-areas, the color features of each sub-area are extracted and counted respectively, the proportion of each color type in the wound area is obtained, and a color distribution histogram is generated.

[0026] In this embodiment, by multi-feature fusion and region growing algorithm, combined with smoothing processing and area difference verification, the accuracy of wound edge contour and area calculation is greatly improved, the deviation of single feature judgment is avoided, multi-source data is integrated to construct a complete three-dimensional model, the three-dimensional form of the wound is intuitively presented, the limitation that the traditional two-dimensional evaluation cannot show the depth and concave is solved; multi-dimensional parameters such as area, depth, texture, color proportion are extracted, the standardization and quantification of wound features are realized, and objective basis is provided for wound severity grading, and accurate data foundation is laid for subsequent dynamic healing curve generation and healing trend prediction.

[0027] In this embodiment, the wound analysis module comprises: The historical data calling unit is configured to receive patient information input by medical staff, including medical record number, name, gender, etc., call historical wound measurement data of the patient, including historical wound three-dimensional model, multi-dimensional wound feature parameters and corresponding measurement time stamp, and construct a wound data time sequence corresponding to the patient; The healing curve generation unit is configured to determine target values of the multi-dimensional wound feature parameters based on the wound data time sequence, and draw an area healing curve, a depth healing curve and a color healing curve of the patient based on the target values; The healing prediction and warning unit is configured to extract multi-dimensional information features of the current patient, perform feature matching on standardized cases in a clinical healing case database, filter reference cases, and generate a corresponding prediction report and risk warning, and generate corresponding intervention suggestions; The report output unit is configured to integrate the dynamic healing curve, the future healing trend prediction curve, the risk warning result and the intervention suggestion into a wound late change prediction report.

[0028] In this embodiment, by accurately associating patient information and historical measurement data, a complete wound data time sequence is constructed, multi-dimensional parameters are converted into intuitive area, depth and color healing curves, and dynamic changes in healing are clearly presented, helping medical staff quickly grasp the healing process, and reference cases are generated by combining a case library and patient feature matching, a healing prediction report and risk warning are generated.

[0029] In one embodiment, the warning step of risk warning in the healing prediction and warning unit comprises: According to the corresponding relationship between the target values and the historical measurement points in the historical wound measurement data, the historical healing rates of each healing type are calculated:

[0030] in, For the first Each historical measurement point corresponds to a healing type. The target value, For the first Each historical measurement point corresponds to a healing type. The target value, For the first Each historical measurement point corresponds to a healing type. healing rate For the first From the first historical measurement point to the... Time interval between historical measurement points; healing type Including: area healing, depth healing, and color healing. ; Based on historical healing rates, calculate the mean and standard deviation of historical healing rates for each healing type:

[0031]

[0032] in, Healing type at the current moment The historical average healing rate, Healing type at the current moment Historical healing rate standard deviation The total number of historical measurement points, the first The historical measurement points are the measurement points at the current moment; Calculate the dynamic risk threshold for each healing type based on the historical mean healing rate and the historical standard deviation of the healing rate:

[0033] in, Healing type at the current moment Dynamic risk threshold, For risk coefficient, ; If the current healing rate If the value is less than the corresponding dynamic risk threshold, a delayed healing warning of the corresponding healing type is generated, and the delayed healing time is calculated:

[0034] in, Healing type at the current moment Delayed healing time, For healing type The goal of healing The target healing time point; The delayed healing warning and the corresponding delayed healing time are jointly warned as risk information.

[0035] In the embodiment, the current healing rate Corresponding less than the corresponding dynamic risk threshold refers to the correspondence of the healing type.

[0036] For example, the wound type is chronic diabetic foot ulcer, and the wound is analyzed as follows: The healing target is: the area healing target: The target healing time point is 120 days, the current time point is 100 days, the total number of historical measurement points is 100, and the time interval between the historical measurement points is 1 day.

[0037] It is known , It is calculated that , , 0.0263 is less than 0.0375, and it is calculated that The warning of delayed healing of the wound area (delayed healing warning of the corresponding healing type) is output, and the warning details of the expected delayed healing of 120 days are attached in the warning.

[0038] The working principle and beneficial effects of the above technical solution are: The dynamic risk model is established by the historical data statistics, the false judgment caused by using the fixed threshold is avoided, the threshold is more adaptive, and the reminding accuracy of the healing risk is improved. The delayed healing time in the risk information provides quantitative prediction, which provides time guidance for early intervention, and further improves the rationality of subsequent intervention.

[0039] In one embodiment, the light supplement module further comprises: According to the collection reference distance of the handheld intelligent measurement terminal and the skin wound area surface, a terminal three-dimensional model is constructed in the three-dimensional space of the wound three-dimensional model; According to the light supplement strategy preset by the light supplement device, the light supplement in the three-dimensional space is simulated, and the first direction vector from the light source center of the virtual light supplement device to each surface point under each light supplement strategy is determined; According to the perpendicular line from the surface point to the normal skin surface around the wound, a second direction vector is constructed; The vector angle of the first direction vector and the second direction vector is calculated and used as the light supplement angle of the corresponding surface point; According to the preset light supplement angle suitability evaluation library, the first suitability value of the light supplement angle is matched and associated with the surface point; The illumination parameter of each surface point under each light supplement strategy is determined; ​According to the color feature and the texture feature of the surface point, a second suitable value of the light parameter is determined based on a light illumination suitability evaluation model, and is associated with the surface point; The first suitable value and the second suitable value associated with the surface point are fused by weighting, to obtain a target suitable value corresponding to the surface point; The target suitable values corresponding to all surface points under the same light compensation strategy are summed to obtain a suitability degree corresponding to the light compensation strategy; The light compensation strategy with the highest suitability degree is applied.

[0040] In this embodiment, when constructing the terminal three-dimensional model, a basic three-dimensional model is first constructed according to the structural parameters of the handheld intelligent measurement terminal, and then the basic three-dimensional model is mapped in the three-dimensional space of the wound three-dimensional model according to the collected reference distance.

[0041] In this embodiment, the light compensation device is a light source system integrated on the handheld intelligent measurement terminal. The light compensation strategy is a preset group of light compensation parameter combinations, including: light compensation device position, illumination angle, light intensity and color temperature.

[0042] In this embodiment, the virtual light compensation device is a virtual model part corresponding to the light compensation device in the terminal three-dimensional model.

[0043] In this embodiment, when determining the first directional vector of the light source center of the virtual light compensation device to each surface point, the light source center of the virtual light compensation device is taken as the starting point of the vector, the surface point is taken as the ending point of the vector, and the distance from the light source center to the surface point is taken as the module length to construct the vector, that is, the first directional vector is obtained.

[0044] In this embodiment, when constructing the second directional vector, the perpendicular and the perpendicular point of the normal skin surface of the wound periphery are taken as the starting point of the vector, the surface point is taken as the ending point of the vector, and the distance from the perpendicular point to the surface point is taken as the module length to construct the vector, that is, the second directional vector is obtained.

[0045] In this embodiment, the preset light compensation angle suitability evaluation library is a pre-established database, which records the suitability degree of viewing the light compensation target under different light compensation angles. The smaller the light compensation angle is, the higher the suitability degree (first suitable value) is.

[0046] In this embodiment, the light parameter is the light intensity and the color temperature.

[0047] In this embodiment, the light illumination suitability evaluation model is a machine learning model for evaluating whether the light parameter is conducive to the clear presentation of the wound feature according to the color feature and the texture feature. The color feature is the color information of the surface point, and the texture feature is the roughness, granularity and wetness of the surface point.

[0048] In this embodiment, the light illumination suitability evaluation model is trained by a plurality of artificial evaluation records of whether the known light illumination parameters are conducive to the clear presentation of wound features according to the wound images labeled with color features and texture features. During the training, the labeled color features, texture features and known light illumination parameters are used as input parameters of the machine learning model, and the artificial evaluation value is used as the output parameter.

[0049] In this embodiment, the second suitability value is an evaluation value of whether the light illumination parameter is conducive to the clear presentation of wound features evaluated by the light illumination suitability evaluation model, and the higher the evaluation value, the more conducive the corresponding light illumination parameter is to the clear presentation of wound features.

[0050] In this embodiment, the weighted fusion refers to assigning dynamic weights to the first suitability value and the second suitability value to obtain a comprehensive suitability degree score (target suitability value).

[0051] In this embodiment, the dynamic of the dynamic weight refers to dynamically adjusting the fusion weights of the first suitability value and the second suitability value according to the difference between the light compensation strategy with the highest suitability degree determined by the system each time and the light compensation strategy with the highest actual suitability degree, so that the light compensation strategy with the highest suitability degree determined subsequently tends to the light compensation strategy with the highest actual suitability degree.

[0052] The working principle and beneficial effects of the above technical solution are: When compensating light for a skin wound, the setting of the light compensation parameter also greatly affects the wound measurement effect, and the more suitable light compensation effect is more conducive to the extraction of subsequent analysis data. Therefore, the present application constructs a terminal three-dimensional model in the three-dimensional space of the wound three-dimensional model, and simulates light compensation in the three-dimensional space according to the light compensation strategy. The first direction vector and the second direction vector are introduced, the influence of the light compensation angle on the suitability degree of wound viewing is quantified (the first suitability value) by calculating the vector angle of the first direction vector and the second direction vector. The light illumination suitability evaluation model is introduced, the second suitability value of the light illumination parameter is determined according to the color features and texture features of the surface points, the first suitability value and the second suitability value associated with the surface points are weighted and fused to obtain the target suitability value, the target suitability values corresponding to all surface points under the same light compensation strategy are summed to obtain the suitability degree of the corresponding light compensation strategy, and the light compensation strategy with the highest suitability degree is applied. The present application realizes fully automatic and intelligent light compensation decision by comprehensively considering the geometric light illumination rationality and medical feature visibility, the dynamic weight mechanism enables the system to continuously calibrate the evaluation standard according to the actual use feedback, realizes self-adaptive optimal light compensation control, improves the light compensation effect, and further improves the subsequent wound analysis accuracy.

[0053] The above merely describes preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art, according to the technical solution and inventive concept of the present application, makes equivalent replacement or change within the technical range disclosed by the present application, should be covered within the protection scope of the present application.

Claims

1. A smart terminal for measuring and analyzing skin wound area, characterized in that, It includes a handheld intelligent measurement terminal, which integrates a wound data acquisition module, a wound model construction module, and a wound analysis module. The wound data acquisition module is used to simultaneously capture wound data information based on depth sensors and optical imaging cameras, process and integrate the acquired wound data information, and generate corresponding three-dimensional point cloud data. The wound model construction module is used to preprocess 3D point cloud data, construct a complete 3D wound model based on 3D point cloud data corresponding to multiple views, and determine the wound edge contour. The wound analysis module is used to retrieve historical wound measurement data of the same patient, generate a dynamic healing curve based on time series, and mark key healing nodes. It also establishes a dynamic risk model through historical data statistics for risk warning and integrates multi-dimensional information to generate a report predicting changes in the wound in the later stage and adjustment suggestions.

2. The intelligent measurement and analysis terminal for skin wound area as described in claim 1, characterized in that, The wound data acquisition module includes: The data acquisition and calibration unit is used to calibrate each component before acquiring skin wound data. After calibration, the laser projection mode and acquisition parameters are determined based on the pre-input operation requirements and wound type information. The data synchronization capture unit is used to synchronously acquire wound area, raw depth data and raw optical data of the patient's wound area based on a determined laser projection mode and acquisition parameters, and to monitor the working synchronization of each component in real time during the acquisition process. The wound data integration unit is used to preprocess the acquired raw depth data and raw optical data, establish the correspondence between the raw optical data and raw depth data, extract wound color feature data by combining projection scale data, and generate three-dimensional point cloud data.

3. The intelligent measurement and analysis terminal for skin wound area as described in claim 2, characterized in that, The wound data acquisition module also includes: The distance between the handheld intelligent measurement terminal and the surface of the skin wound area is extracted based on the depth sensor, and the imaging distance of the optical imaging camera is extracted at the same time. The imaging distance is then compared and calibrated with the acquisition reference distance. The matching degree between the calibrated acquisition reference distance and the preset laser projection adaptation threshold is calculated, and the current distance is determined to be within the effective measurement range based on the calculation result. Establish the matching relationship between the acquisition reference distance and the grid density, combine the preset matching distance parameter reference value, determine the matching distance parameter of each component, and set the acquisition parameters of each component based on the matching distance parameter.

4. The intelligent measurement and analysis terminal for skin wound area as described in claim 3, characterized in that, The data synchronization capture unit also includes: Based on the determined laser projection mode, a grid-like projection scale and a corresponding skin damage color comparison map are projected onto the patient's wound and surrounding normal skin area. The corresponding components are controlled to scan and collect data on the patient's wound area according to the preset acquisition parameters. The laser projection component continuously projects a grid-like projection scale, counts the number of complete grids and half grids covering the wound, obtains preliminary wound area data in real time, and compares the actual color of the wound with the projected standard color comparison map to determine the wound color type. A depth sensor collects depth information at different locations of the wound in real time, while an optical imaging camera simultaneously collects image data of the patient's wound area, acquiring raw depth data and raw optical data simultaneously.

5. The intelligent measurement and analysis terminal for skin wound area as described in claim 4, characterized in that, The wound model building module includes: The edge contour determination unit is used to extract the depth difference, color difference, and texture difference between the wound area and the surrounding normal skin area in the preprocessed 3D point cloud data, and to construct a boundary judgment index based on multi-feature fusion. Using preset normal skin pixels as seed points, and combining boundary judgment indicators, the normal skin area is gradually grown and expanded to determine the growth boundary of the area not covered by growth. The growth boundary is smoothed to generate continuous wound edge contour data. At the same time, the area of ​​the region enclosed by the contour is calculated and the difference is calculated with the wound area calculated by the wound data acquisition module. If the difference exceeds the preset area deviation threshold, the three-dimensional point cloud data preprocessing operation is performed again. The model output unit is used to construct a complete 3D model of the wound based on the determined wound edge contour, combined with the original optical data and the original depth data, and synchronously transmits the constructed complete 3D model of the wound and the wound edge contour data to the wound analysis module.

6. The intelligent measurement and analysis terminal for skin wound area as described in claim 5, characterized in that, The wound model building module also includes a wound feature quantization unit, which is used to extract features from the complete 3D wound model and generate multi-dimensional wound feature parameters. The edge contour is generated based on the wound edge contour determination unit, and the actual area of ​​the wound is obtained based on the coordinates of the contour discrete points. Traverse each surface point in the 3D wound model, calculate the vertical distance between each surface point and the normal skin surface around the wound, record the maximum distance as the maximum depression depth of the wound, and calculate the average distance of all depression points as the average depression depth. Extract the gray-level co-occurrence matrix of the wound surface model, calculate the texture feature parameters such as contrast, correlation, and energy of the gray-level co-occurrence matrix, and quantitatively characterize the roughness and tissue uniformity of the wound surface. Based on the wound color feature data, the wound area is divided into multiple sub-regions. The color features of each sub-region are extracted and statistically analyzed to obtain the proportion of each color type in the wound area and generate a color distribution histogram.

7. The intelligent measurement and analysis terminal for skin wound area as described in claim 6, characterized in that, The wound analysis module includes: The historical data retrieval unit is used to receive patient information input by medical staff, retrieve the patient's historical wound measurement data, including historical wound 3D models, multi-dimensional wound feature parameters and corresponding measurement timestamps, and construct the wound data time series corresponding to the patient. The healing curve generation unit is used to determine the target values ​​of multi-dimensional wound feature parameters based on the wound data time series, and to draw the patient's area healing curve, depth healing curve and color healing curve based on the target values; The healing prediction and early warning unit is used to extract multi-dimensional information features of the current patient, perform feature matching on standardized cases in the clinical healing case database, screen reference cases, generate corresponding prediction reports, provide risk warnings, and generate corresponding intervention suggestions. The report output unit integrates the dynamic healing curve, the future healing trend prediction curve, the risk warning results, and the intervention suggestions into a wound post-change prediction report.

8. The intelligent measurement and analysis terminal for skin wound area as described in claim 7, characterized in that, The risk warning steps in the healing prediction and early warning unit include: Based on the correspondence between target values ​​and historical measurement points in historical wound measurement data, the historical healing rate for each healing type is calculated; Based on the historical healing rate, calculate the mean and standard deviation of the historical healing rate for each healing type. Calculate the dynamic risk threshold for each healing type based on the historical mean healing rate and the historical standard deviation of the healing rate; If the current healing rate is less than the corresponding dynamic risk threshold, a delayed healing warning of the corresponding healing type is generated, and the delayed healing time is calculated. The delayed healing warning and the corresponding delayed healing time will be presented together as risk information.

9. The intelligent measurement and analysis terminal for skin wound area as described in claim 1, characterized in that, The handheld intelligent measurement terminal also integrates a supplementary lighting module and a power module. The supplementary lighting module automatically adjusts the brightness according to the ambient light intensity, and the power module is a rechargeable lithium battery.

10. The intelligent measurement and analysis terminal for skin wound area as described in claim 9, characterized in that, The supplementary lighting module also includes: Based on the acquisition reference distance between the handheld intelligent measurement terminal and the surface of the skin wound area, a three-dimensional model of the terminal is constructed in the three-dimensional space of the wound three-dimensional model. Based on the pre-set supplementary lighting strategy of the supplementary lighting device, supplementary lighting is simulated in three-dimensional space, and the first direction vector from the light source center of the virtual supplementary lighting device to each surface point is determined under each supplementary lighting strategy. Construct a second direction vector based on the perpendicular lines from the surface point to the normal skin surface surrounding the wound; Calculate the vector angle between the first direction vector and the second direction vector, and use it as the supplementary lighting angle for the corresponding surface point; Based on the preset supplementary lighting angle suitability evaluation library, the first suitable value of the supplementary lighting angle is matched and associated with the surface point; Determine the illumination parameters for each surface point under each supplementary lighting strategy; Based on the illumination suitability evaluation model, a second suitable value of the illumination parameter is determined according to the color and texture features of the surface points and associated with the surface points. The illumination suitability evaluation model is trained by multiple evaluation records of human evaluation of whether the known illumination parameters are conducive to the clear presentation of wound features based on wound images labeled with color and texture features. During training, the labeled color features, texture features and known illumination parameters are used as input parameters of the machine learning model, and the human evaluation value is used as the output parameter. The first and second suitability values ​​associated with the surface points are weighted and fused to obtain the target suitability value corresponding to the surface points. Summing the target suitability values ​​corresponding to all surface points under the same supplementary lighting strategy yields the suitability of the corresponding supplementary lighting strategy. Apply the most suitable supplemental lighting strategy.

Images