Method and system for deformation visual monitoring of the pressurization process of a wound re-warming chamber

Abstract

The application relates to a deformation visualization monitoring method and system for a wound rewarming cabin pressure process. The method comprises the following steps: acquiring baseline imaging data of a wound edge safety area according to the optical geometric relationship between a pose calibration camera and a transparent airbag; controlling the wound rewarming cabin to testively pressurize a wound; calculating a surface displacement strain time sequence field of the wound according to continuous imaging data of the wound edge safety area; controlling the wound rewarming cabin to stepwise pressurize the wound according to the surface displacement strain time sequence field, so as to obtain wound tissue compliance data; controlling the wound rewarming cabin to perform pressure-temperature coupling circulation on the wound according to the wound tissue compliance data, so as to obtain wound tissue maintenance data; performing pressure-temperature coupling perturbation correction on the wound according to the wound tissue maintenance data, so as to obtain wound tissue stability data; and generating deformation visualization data of the wound rewarming cabin. The method can effectively improve the success rate of treatment of wounded persons in cold region activities.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic control, in particular to a deformation visualization monitoring method and system for a wound rewarming cabin pressurization process. BACKGROUND

[0002] When activities are carried out in cold regions, the activity environment is extremely harsh, and extreme conditions such as low temperature, strong wind, and snowstorms occur frequently. After a wounded person is injured, the body is prone to hypothermia due to low temperature, which leads to slow wound healing, abnormal blood clotting, and various complications, greatly increasing the difficulty of treatment and mortality. Traditional vital sign monitoring equipment, such as ordinary life monitoring bracelets, not only has greatly discounted accuracy of monitoring data in a low-temperature environment, but also has relatively single functions, and thus cannot comprehensively and real-timely monitor key vital signs of a wounded person, and cannot provide timely and effective data support for treatment. Meanwhile, existing wound rewarming equipment often operates independently, lacks effective linkage with vital sign monitoring links, and cannot timely adjust rewarming strategies according to dynamic changes of vital signs of a wounded person, leading to low success rate of treatment of wounded persons in cold region activities. SUMMARY

[0003] Therefore, it is necessary to provide a deformation visualization monitoring method and system for a wound rewarming cabin pressurization process, which can effectively improve the success rate of treatment of wounded persons in cold region activities.

[0004] In a first aspect, the present application provides a deformation visualization monitoring method for a wound rewarming cabin pressurization process, comprising:

[0005] In a case where the wound rewarming cabin is not closed, baseline imaging data of a wound edge safety area is acquired according to an optical geometric relationship between a pose calibration camera and a transparent airbag;

[0006] The wound rewarming cabin is controlled to perform test pressurization on a wound, and surface displacement strain time series of the wound is calculated according to continuous imaging data of the wound edge safety area;

[0007] The wound rewarming cabin is controlled to perform stepwise pressurization on the wound according to the surface displacement strain time series, and wound tissue compliance data is obtained;

[0008] The wound rewarming cabin is controlled to perform pressure-temperature coupling circulation on the wound according to the wound tissue compliance data, and wound tissue maintenance data is obtained;

[0009] The wound is subjected to pressure-temperature coupling perturbation correction according to the wound tissue maintenance data, and wound tissue stability data is obtained; and the wound tissue stability data is used to generate deformation visualization data of the wound rewarming cabin.

[0010] In a second aspect, the application also provides a deformation visualization monitoring system for a wound re-warming cabin pressurization process, comprising a terminal and a computer device;

[0011] The computer device is configured to, in the case that the wound re-warming cabin is not closed, acquire baseline imaging data of the safe area of the wound margin according to the optical geometric relationship between the pose calibration camera and the transparent airbag; and the terminal is configured to detect the closing condition of the wound re-warming cabin.

[0012] The computer device is configured to control the wound re-warming cabin to perform test pressurization on the wound, and calculate a surface displacement strain time series field of the wound according to the continuous imaging data of the safe area of the wound margin.

[0013] The computer device is configured to control the wound re-warming cabin to perform stepwise pressurization on the wound according to the surface displacement strain time series field, and obtain wound tissue compliance data.

[0014] The computer device is configured to control the wound re-warming cabin to perform pressure-temperature coupling circulation on the wound according to the wound tissue compliance data, and obtain wound tissue maintenance data.

[0015] The computer device is configured to perform pressure-temperature coupling perturbation correction on the wound according to the wound tissue maintenance data, and obtain wound tissue stability data; and the wound tissue stability data is used to generate deformation visualization data of the wound re-warming cabin.

[0016] The above-mentioned deformation visualization monitoring method and system for a wound re-warming cabin pressurization process first establish the optical geometric relationship between the camera and the transparent airbag by pose calibration in the case that the wound re-warming cabin is not closed, acquire baseline imaging data of the safe area of the wound margin, and ensure non-invasive and safe imaging of the wound; then perform test pressurization on the wound under control, and calculate a surface displacement strain time series field in combination with continuous imaging results, so that the small deformation of the wound tissue can be dynamically and accurately captured; further stepwise pressurization testing is performed to obtain tissue compliance data, which provides a scientific basis for evaluating the biomechanical response of the wound under different external forces; then pressure-temperature coupling circulation testing is combined to obtain tissue maintenance data, which reflects the functional recovery ability of the wound tissue under comprehensive stress conditions from multiple dimensions; finally, pressure-temperature coupling perturbation correction is performed to generate data visualization results that can directly show the stability of the wound. Not only does this realize systematic and quantitative evaluation of key parameters such as strain characteristics, compliance and stability of the wound tissue, but also provides scientific and reliable evidence for formulating individualized treatment strategies in clinical practice, greatly improving the safety, pertinence and controllability of the wound treatment process, and effectively improving the success rate of treatment of wounded personnel in cold regions. BRIEF DESCRIPTION OF DRAWINGS

[0017] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is an application environment diagram of the deformation visualization monitoring method during the pressurization process of the wound rewarming chamber in one embodiment;

[0019] Figure 2 This is a flowchart illustrating a deformation visualization monitoring method for the pressurization process of a wound rewarming chamber in one embodiment.

[0020] Figure 3 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0022] This application provides a method for visually monitoring deformation during the pressurization process of a wound rewarming chamber, which can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located in the cloud or on other network servers. Server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0023] In one exemplary embodiment, such as Figure 2 As shown, a method for visually monitoring deformation during the pressurization process of a wound rewarming chamber is provided, which can be applied to... Figure 1 Taking the server in the example, the explanation includes the following steps 202 to 210. Wherein:

[0024] Step 202: With the wound rewarming chamber not closed, baseline imaging data of the safe area of ​​the wound edge is obtained based on the optical geometric relationship between the pose calibration camera and the transparent airbag.

[0025] Step 204: Control the wound rewarming chamber to apply test pressure to the wound, and calculate the surface displacement strain time series field of the wound based on the continuous imaging data of the safe area of ​​the wound edge.

[0026] Step 206, according to the surface displacement strain timing field, control the wound rewarming cabin to step up pressure on the wound, get wound tissue compliance data.

[0027] Step 208, according to the wound tissue compliance data, control the wound rewarming cabin to carry out pressure-temperature coupling cycle on the wound, get wound tissue maintenance data.

[0028] Step 210, according to the wound tissue maintenance data, carry out pressure-temperature coupling perturbation correction on the wound, get wound tissue stability data.

[0029] Among them, the wound rewarming cabin is a closed device for implementing pressure control and temperature control treatment on the wound surface, including transparent air bag (pressure medium), partition heating / temperature control module, pressure and temperature sensor, imaging window and camera support, etc., its core responsibility is to apply programmable pressure-temperature trajectory to the wound according to the control strategy.

[0030] Among them, the pose calibration camera is an imaging unit and its calibration process for obtaining the image of the wound edge area and participating in geometric calibration; by calibrating the internal and external parameters of the camera, the installation angle and the relative position of the imaging window / air bag, the one-to-one correspondence of "pixel coordinates-wound inherent coordinates" is established.

[0031] Among them, the transparent air bag is a transparent elastic cavity covering the wound surface, which is a pressure applying and external isolation medium; its transparency allows the camera to observe the wound edge, and the elastic deformation realizes controllable pressure; the geometric / optical properties (curvature / refraction) are compensated by algorithm after calibration to avoid the influence of imaging distortion on measurement.

[0032] Among them, the optical geometric relationship is the geometric and refractive mapping of the whole imaging link of camera-imaging window-transparent air bag-wound surface, including internal / external parameters, distortion model and cross-medium (air / material) refraction model; the relationship is used to restore the original image to the wound inherent coordinates, and eliminate the system error introduced by the carrier and the medium.

[0033] Among them, the wound edge safety area is the wound edge near neighbor area (ROI) determined by doctors / algorithm together in the baseline stage, which is suitable for measurement and force; it excludes high-risk / low-confidence areas such as necrosis, exudation, reflection and edge tearing, and is used as the unified analysis range for subsequent displacement-strain calculation, threshold determination and visualization.

[0034] Among them, the baseline imaging data is a set of wound edge images collected and denoised after optical geometric calibration before the rewarming cabin is closed or pressurized, which records the reference state of "zero pressure zero heat", and contains ROI mask and coordinate mapping.

[0035] Wherein, the test pressure is a small, controllable pressure rise (usually slow, short, and constant temperature) applied in the early stage of treatment to detect tissue response. The purpose is to induce a measurable deformation response within the safety boundary.

[0036] Wherein, the continuous imaging data is a time series of images collected from the safety area of the wound margin at a fixed frame rate during the test pressure and subsequent processes.

[0037] Wherein, the surface displacement strain time series field is a displacement vector field and a corresponding surface strain field sequence calculated in the wound margin intrinsic coordinate system, which reflects the true tissue response caused by pressure after correcting medium distortion and non-pressure components.

[0038] Wherein, the stepwise pressure is a stepwise pressure protocol (equal step, pressure holding at each level and stability judgment) implemented to obtain steady state and compliance; the steady state statistical points of each step are used to fit the pressure-strain relationship and identify the transition interval from "easy compression to difficult compression", avoiding the viscoelastic deviation caused by continuous climbing.

[0039] Wherein, the wound tissue compliance data is a set of tissue pressure response information obtained by stepwise pressure sampling and fitting, including the structured description of each level steady state characteristics, fitting curve and (optional) transition interval evidence; it is used to develop the initial target pressure zone and subsequent pressure-temperature coupling strategy.

[0040] Wherein, the pressure-temperature coupling cycle is a closed-loop execution process that adjusts pressure and partition temperature in conjunction and monitors deformation and quality indicators in real time under the premise of ensuring safe deformation envelope; its output is the uniformity of the maintenance period, strain uniformity and risk event record.

[0041] Wherein, the wound tissue maintenance data is a set of operation data collected and summarized in the stable stage of the pressure-temperature coupling cycle, reflecting the effect and boundary margin of long-term maintenance within the target zone; it provides the basis for perturbation correction and long-term self-adaptation (such as whether the hot spot suppression is effective, whether it is close to the risk of penetration).

[0042] Wherein, the pressure-temperature coupling perturbation correction is a process of adding low-amplitude and low-frequency pressure / power perturbations near the maintenance target, and correcting the target setting and partition weight by observing the changes in tissue sensitivity; it is used to verify whether the steady state is truly stable, and to make small step corrections for potential drift (edema, body position change).

[0043] Wherein, the wound tissue stability data is a set of final treatment settings and stability evidence formed after the convergence of the perturbation correction, including the remaining safety margin, sensitivity spectrum, partition weight reduction / shielding suggestion and abnormal rollback rule.

[0044] The deformation visualization data is a visualization output (strain heat map, displacement vector superposition, wound edge ring curve, risk mask, and event timeline, etc.) generated based on the final surface displacement-strain time series, steady-state window, and stability data, and is attached with a timestamp and a set parameter watermark for clinical interpretation, efficacy tracking, and audit traces.

[0045] Specifically, in the case of the wound re-warming cabin not being closed, the calibrated target (or using the natural texture of the wound edge) is placed to calibrate the optical geometric relationship of the pose calibration camera-transparency airbag, which simultaneously calibrates the camera internal and external parameters, lens distortion, and airbag / window refraction model, and establishes the mapping of image to wound inherent coordinates; under the conditions of fixed ambient light and exposure, ≥N baseline images are collected for time domain denoising, and the "wound edge safe area" ROI is automatically / semi-automatically outlined and fixed; the baseline imaging data package is generated, which contains the inherent coordinate grid, refraction / distortion correction parameters, and ROI mask.

[0046] The server 104 controls the wound re-warming cabin to start a small "probing pressurization" (slow rise, constant temperature) to test the wound, and continuously images the wound edge safe area at a fixed frame rate during the test pressurization; the relevant matching (DIC) and optical flow are calculated in parallel in the inherent coordinate system, and then the "unpressurized counterfactual" trajectory is generated to separate the non-pressurization component, and the sub-pixel displacement time series is obtained by fusion; the displacement field is subjected to topological-physical consistent strain reconstruction (boundary conservation, radial compliance monotonicity), forming the surface displacement-strain time series field, and the quality indicators (trend leveling, low fluctuation, hotspot non-expansion) are output synchronously.

[0047] The server 104 controls the wound re-warming cabin to perform a stepwise pressurization protocol (equal step + pressure holding window) to perform stepwise pressurization, and at the end of each step, the stability is automatically judged according to the three criteria of "small slope, low variance, and non-expanding hot spot", and the representative strain / displacement statistics of the step are extracted; the individualized compliance curve of pressure-strain is fitted across the steps, and the transition interval from "easy compression to difficult compression" is identified, forming the wound tissue compliance data, which contains the steady-state evidence of the step, the fitting residual, and the candidate safe / effective pressure band.

[0048] Based on the wound tissue compliance data (or the generated pressure-temperature optimization data), the wound re-warming cabin is controlled to perform pressure-temperature coupling circulation on the wound, and during the execution of the pressure-temperature coupling circulation, the pressure is adjusted in small steps within the target band, and the temperature is uniformly heated according to the partition PWM; the controller compares the safety deformation envelope (SDE) and the quality indicators in real time, and automatically reduces the pressure / power if it approaches the boundary; the strain uniformity, hotspot suppression effect, temperature consistency, and event log during the maintenance period are continuously recorded, and are summarized into wound tissue maintenance data.

[0049] Superimpose low-amplitude low-frequency pressure perturbations (or slight zonal power perturbations) near the wound tissue maintenance data, and observe the sensitivity and stability of the strain time sequence to the perturbations: if the hotspot sensitivity increases or the connectivity tends to be through, the target pressure / shielding corresponding partition is adjusted downward and rewarmed; cycle iteration until the sensitivity tends to be stable, the hotspot does not expand, and the boundary margin is sufficient, and the wound tissue stability data is obtained, which contains the final setting, zonal weight reduction / shielding suggestion, remaining safety margin and sensitivity spectrum.

[0050] Synchronize the surface displacement-strain time sequence, steady-state window and stability data in the wound tissue stability data with the uniform time axis and the inherent coordinates of the wound edge, and after determining the final target pressure / temperature and the corresponding steady-state interval, superimpose the displacement vector and the wound edge ring curve on the strain heat map as the main layer, and superimpose the safety deformation envelope, the risk of penetration and the hotspot mask, and the key event markers, while the steady-state evidence, the two-flow consistency residual and the uncertainty prompt are solidified as the evidence layer of each frame. After standardization and scale annotation, output the dynamic picture and the static snapshot (including timestamp and setting parameter watermark) which can be played back and archived, forming the rewarming chamber deformation visualization data which can be used for clinical interpretation, treatment tracking and audit traces.

[0051] In the above-mentioned deformation visualization monitoring method for wound rewarming chamber pressurization process, by using pose calibration to establish the optical geometric relationship between the camera and the transparent airbag under the condition that the wound rewarming chamber is not closed, the baseline imaging data of the wound edge safety area is obtained to ensure non-invasive and safe imaging of the wound; then the wound is tested under control and the surface displacement-strain time sequence field is calculated combined with continuous imaging results, so that the small deformation of the wound tissue can be dynamically and accurately captured; on this basis, further implement the stepwise pressure test to obtain the compliance data of the tissue, which provides a scientific basis for evaluating the biomechanical response of the wound under different external forces; then combined with the pressure-temperature coupling cycle test, the tissue maintenance data is obtained, which reflects the functional recovery ability of the wound tissue under comprehensive stress from multiple dimensions; finally, through pressure-temperature coupling perturbation correction, the data visualization result which can directly show the wound stability is generated. Not only realizes the systematic and quantitative evaluation of key parameters such as strain characteristics, compliance and stability of wound tissue, but also provides scientific and reliable basis for clinical development of individualized treatment strategies, greatly improves the safety, pertinence and controllability of wound treatment process, and can effectively improve the success rate of treatment of wounded personnel in cold region activities.

[0052] In an exemplary embodiment, according to the wound tissue compliance data, the wound rewarming chamber controls the wound to perform pressure-temperature coupling cycle to obtain wound tissue maintenance data, including steps 302 to 308. Among them:

[0053] Step 302, nonlinear fitting is performed on the wound tissue compliance data to obtain wound tissue fitting data.

[0054] Step 304, a wound treatment double-threshold joint analysis is performed according to the wound tissue fitting data to obtain wound pressure-temperature optimization data.

[0055] Step 306, a wound pressure-temperature cycle dynamic analysis is performed on the wound with reference to the wound pressure-temperature optimization data to obtain wound pressure-temperature limit data.

[0056] Step 308, according to the wound pressure-temperature optimization data and the wound pressure-temperature limit data, the wound is controlled to perform pressure-temperature coupling cycle in the wound rewarming cabin to obtain wound tissue maintenance data.

[0057] Among them, the nonlinear fitting is to first do steady-state discrimination and de-lag shaping on the compliance samples obtained by "step-up pressure", and then perform curve fitting under the constraints of monotonicity, amplitude limitation, fast first and slow later, and "as few inflection points as possible", so as to obtain individualized curve which can truly reflect the "pressure-deformation" relationship, rather than simple least squares empirical curve.

[0058] Among them, the wound tissue fitting data is a structured result set generated by the above nonlinear fitting process and directly called by the subsequent steps, at least including: fitting results for representing individualized "pressure-deformation" relationship, error / uncertainty information matched therewith, and evidence index related to sample steady state.

[0059] Among them, the treatment double-threshold joint analysis is a joint determination process based only on "wound tissue fitting data": on the one hand, it identifies the effective hemostatic threshold (EHST, interval definition) of the compliance curve from "easy compression" to "difficult compression", and on the other hand, it generates a spatio-temporal joint safety deformation envelope (SDE) according to the deformation space statistics and uncertainty boundary of the fitting data; the two thresholds are used as hard constraints to jointly limit the treatment feasible region, avoiding the empirical deviation of single-point threshold.

[0060] Among them, the wound pressure-temperature optimization data is an execution setting set obtained by multi-objective (hemostasis effect, deformation uniformity, energy / temperature rise constraint) robust optimization under the constraints of EHST and SDE, usually including target pressure, target temperature, partition heating weight and maintenance period perturbation strategy; this data converts "what can be done" into "how should it be done", and is used to drive the subsequent closed loop and simulation.

[0061] Among them, the pressure-temperature cycle dynamic analysis is referenced to the wound pressure-temperature optimization data, combined with the uncertainty in the fitting data, to perform time domain scanning and boundary tracking on the cycle templates such as pressure increase, constant pressure temperature rise, maintenance period perturbation and pressure decrease, to identify when to touch the SDE boundary, hotspot and approach to approach events, and to quantify the allowed rate, residence time and safety margin process.

[0062] Among them, the wound pressure-temperature limit data is the executable constraint list and guard rule output by the pressure-temperature cycle dynamic analysis, which at least specifies the acceptable maximum / minimum rate of each stage, minimum residence time, perturbation amplitude / frequency range, partition weight or shielding suggestion and prohibited pressure-temperature combination; Its role is to act as a "hard guardrail" in actual closed-loop control to ensure that the optimization goal is implemented within the safe and feasible domain.

[0063] Specifically, according to the "pressure-steady-state evidence" of each level of the step-up pressure in the wound tissue compliance data, the transition samples are removed and the hysteresis section is reshaped to make the samples equivalent to path-independent steady-state points; Then, under the shape priors such as monotonicity, amplitude limitation, fast first and slow later, and the constraint of minimum turning section (sparse inflection point), the data processed in the previous step are nonlinearly fitted, outputting an individualized pressure-deformation compliance curve, and generating an error band, a steady-state window index and a structured field for subsequent judgment, thus forming the wound tissue fitting data.

[0064] The turning interval of the wound tissue fitting data is robustly identified, and the interval "from easy compression to difficult compression" is defined as the effective hemostasis threshold (EHST); and the deformation space statistics and uncertainty boundary in the wound tissue fitting data are used to generate a spatio-temporal joint safety deformation envelope (SDE) to exclude combinations that may have the risk of sheeting or penetration; Finally, taking EHST and SDE as hard constraints, a multi-objective robust optimization (efficacy improvement, deformation uniformity, energy constraint) is constructed to obtain the target pressure, target temperature, partition heating weight and perturbation strategy, forming the wound pressure-temperature optimization data.

[0065] The wound pressure-temperature optimization data is mapped to the "pressure-temperature feasible channel" on the time axis, and according to the uncertainty range given by the wound pressure-temperature optimization data, the cycle templates such as pressure increase, constant pressure temperature rise, maintenance period perturbation and pressure decrease are scanned in an adversarial counterfactual manner, and the earliest time and conditions for triggering SDE boundary, hotspot and approach to approach events are tracked; On this basis, the feasible channel is safely tightened and the rhythm is delimited, and the allowed maximum / minimum rate of each stage, minimum residence time, perturbation amplitude / frequency range and prohibited pressure-temperature combination list are output to form the wound pressure-temperature limit data.

[0066] The server 104 sets a target according to the wound pressure-temperature limit data and performs a closed loop according to the limit data, such as small-step adjustment of pressure within the target band, temperature equalization according to the partition, online monitoring of the deformation timing and the distance from the safety envelope, immediate pressure reduction, power reduction, or shielding of the partition according to the limit rules when a trigger event is encountered, and recording of the rollback, etc.; in the maintenance period, low-amplitude perturbations are periodically applied to verify the perfusion / deformation stability and fine-tune the rhythm parameters, and finally the running indicators such as target achievement, hot spot suppression effect, boundary margin, perturbation response, and abnormal handling log are settled, and the wound tissue maintenance data is summarized.

[0067] In this embodiment, the transition and hysteresis deviation is first removed by nonlinear fitting to obtain stable individual response; and a double threshold (therapeutic EHST + safety SDE) is used as a hard constraint to generate pressure-temperature optimization data, which takes into account hemostasis and tissue safety from the source; dynamic cycle analysis gives the combination of lift rate, minimum residence, and disabled in advance, reducing the risk of edge touch, penetration, and overheating; finally, the "optimization + limit" contract is executed in a closed loop and is self-proven in a stable state by perturbation, forming traceable maintenance data. It can achieve faster threshold, more stable maintenance, lower concurrency, and higher repeatability / auditability.

[0068] In one exemplary embodiment, according to the wound tissue fitting data, a wound is treated by double-threshold joint analysis to obtain wound pressure-temperature optimization data, including steps 402 to 410. Among them:

[0069] Step 402, the wound tissue fitting data is structurally parsed to obtain compliance curve data, deformation field statistical data, fitting parameter error range, and pressure retention stable state evidence.

[0070] Step 404, the pressure retention stable state evidence is windowed and screened to obtain a stable state data subset.

[0071] Step 406, according to the compliance curve data and the fitting parameter error range, the turning interval of the wound is analyzed to obtain effective hemostatic threshold pressure data.

[0072] Step 408, according to the deformation field statistical data, the fitting parameter error range, and the stable state data subset, the wound is analyzed for spatiotemporal connectivity stability to obtain safety deformation envelope data.

[0073] Step 410, the effective hemostatic threshold pressure data and the safety deformation envelope data are processed by robust optimization to obtain wound pressure-temperature optimization data.

[0074] Among them, the structured analysis is the process of regularizing and disassembling the "wound tissue fitting data" under the inherent coordinates of the wound and the unified time axis. It converts the original / intermediate results into standard views that can be directly called by subsequent algorithms. It usually synchronously outputs four types of core elements: compliance curve data, deformation field statistical data, fitting parameter error range, and pressure maintaining steady state evidence. It establishes the index relationship between them and the metadata such as step pressure, timestamp, ROI, etc. to avoid repeated calculation and ensure traceability and consistency.

[0075] Among them, the compliance curve data is an individualized function / segmented representation reflecting the pressure-deformation (or strain) relationship and its sampling point mapping. It is derived from the steady state statistical points of each level of the ladder and fitted by shape prior constraint. It is used to identify the transition behavior from "easy compression" to "difficult compression" and is the direct basis for determining the effective hemostasis threshold and formulating the target pressure zone.

[0076] Among them, the deformation field statistical data is a spatial-temporal statistic obtained by summarizing the surface displacement-strain time field in the steady state window of each step. It typically includes the circumferential average strain, the maximum principal strain, the number / area / max span / occurrence duration of hot spots, and the minimum distance from the wound edge, etc. It is used to evaluate whether it is in pieces or not and its evolution trend.

[0077] Among them, the fitting parameter error range is the uncertainty bound obtained after residual analysis, cross-validation or scenario sampling of the compliance fitting and related mapping. It is used to set a robust threshold and safety margin for threshold identification, connectivity analysis and optimization solution, to avoid misjudgment and out-of-bound caused by point estimation.

[0078] Among them, the pressure maintaining steady state evidence is a joint criterion for the "trend flattening" of the strain curve in the short window of the pressure maintaining phase, the "continuous convergence" of the fluctuation degree, and the "spatial non-expansion" of the hot spot, as well as its timestamp record. If necessary, it is supplemented by pressure stability and visual quality safeguard, which is used to confirm which frames / interval can be used as reliable samples for fitting and safety analysis.

[0079] Among them, the steady state data subset is the time period and its corresponding pressure range and statistical quantity set retained after screening by the pressure maintaining steady state evidence. It is the only time window source for subsequent turning interval identification (EHST) and spatio-temporal connectivity stability analysis (SDE), ensuring that all determinations are based on high reliable data.

[0080] Among them, the turning interval is the pressure section on the compliance curve that stably transitions from "easy compression" to "difficult compression". It is usually detected by the joint detection of significant slope decrease and curvature transition, combined with uncertainty checking and steady state backtracking confirmation.

[0081] Among them, the effective hemostasis threshold pressure data is the pressure range result output on the basis of the turning interval after robustness and repeatability checking, with boundary confidence and source step index, used to constrain subsequent optimization and control to make treatment "sufficient to stop bleeding" while avoiding excessive compression.

[0082] Among them, the spatiotemporal connectivity stability analysis is a joint process of adaptive threshold, high strain mask generation, directional penetration / seepage test, morphological counterfactual robustness test and persistence analysis on the strain field in the steady state window, to quantify whether the hot spots are patchy, whether they tend to be penetrated, and their sensitivity to uncertainty.

[0083] Among them, the safe deformation envelope data is a spatial safety mask, risk mask and conservative buffer zone output according to the pressure level, and can be accompanied by the result set of partitioned weight reduction / shielding recommendations, as a "hard constraint" for optimization and control, to ensure that the pressure and temperature strategy always operates within the safe and feasible domain.

[0084] Among them, the robust optimization process is a multi-objective solving process under the double constraints of effective hemostasis threshold and safe deformation envelope, combined with uncertainty bounding, to improve efficacy and deformation uniformity, and to suppress risk and energy load.

[0085] Specifically, the wound tissue fitting data is structurally analyzed under the inherent coordinates of the wound edge and the unified time axis, to obtain compliance curve data (function / segmented representation of pressure-deformation and corresponding sampling points), deformation field statistical data (ring average strain, maximum principal strain, hot spot connected domain area / span / duration, etc. according to step / time), fitting parameter error range (uncertainty interval / envelope based on fitting residual and cross-validation), and pressure stabilization evidence (trend leveling, fluctuation convergence, hot spot non-expansion determination result and its timestamp, pressure level, etc. metadata).

[0086] According to the "steady state three criteria" (strain increment leveling, short window fluctuation convergence, hot spot non-expansion), the steady state evidence of each pressure holding interval is automatically judged and denoised, and the frame segment containing abnormal residual or occlusion is removed, and the passers are labeled with steady state label and confidence; The data obtained will be aggregated with the corresponding pressure level, deformation statistics and uncertainty annotation of the frame segment that passes the stability judgment, to form a steady state data subset.

[0087] In the unified time axis and the inherent coordinate of the wound edge, the sampling points corresponding to the steady-state data subset are mapped onto the compliance curve data, and multi-scale piecewise fitting and smoothing are performed according to the shape priors of the minimum inflection point and the monotone amplitude limiting. Then, the candidate transition zone with the joint criterion of slope attenuation and curvature transition is screened out from the "easy compression to difficult compression". For each candidate zone, uncertainty checking (leave-one-cross-validation, neighborhood consistency and abnormal step elimination) is performed combined with the fitting parameter error range, and the persistence and repeatability of the start and end positions of the transition zone are verified by the steady-state evidence, so as to compress the effective hemostatic threshold into an interval. Finally, the effective hemostatic threshold pressure data is output, in which the transition interval, its confidence label and the source step index are given as the direct input for subsequent safety envelope analysis and robust optimization.

[0088] In the inherent coordinate system of the wound edge, adaptive segmented threshold and uncertainty envelope are implemented on the strain field in the steady-state data subset according to the deformation field statistical data and the fitting parameter error range, and a high-strain mask sequence is generated. At the same time, under the constraint of the wound-peripheral boundary, the sequence is subjected to directional penetration / flowing judgment and the minimum fracture size is evaluated, and the time sequence trajectories of the number, area, maximum span and residence time of the connected domain are recorded simultaneously. Further, the mask is subjected to morphological dilation / erosion and counterfactual robustness test of random pose disturbance within the uncertainty range, and the accidental short-term hotspots are eliminated by the persistence analysis, to obtain the penetration risk map, the robustness score map and the residence expansion evidence map. Finally, the above-mentioned layers are subjected to multi-layer fusion and cascade threshold according to the pressure level, and the safety deformation envelope data with index is output, which clearly shows the spatial safety mask, risk mask and conservative buffer zone, and recommends the weight reduction / shielding partition.

[0089] Taking the effective hemostatic threshold pressure range (EHST) and the safety deformation envelope data (SDE) as hard constraints, and combining the fitting parameter error range, a distributed robust multi-objective optimization model is constructed. Under the cooperation of scenario sampling and constraint relaxation penalty, the therapeutic effect gain and deformation uniformity / temperature balance are simultaneously improved, the hotspot risk and energy load are inhibited, and the infeasible solution is projected back to the feasible region. The decision variables include target pressure, target temperature, partition heating weight, and perturbation amplitude and frequency in the maintenance period. After solving, the wound pressure and temperature optimization data can be directly issued, which clearly shows the set values, partition weight reduction / shielding suggestions, minimum boundary margin for SDE, and scenario coverage rate, etc. executable and auditable indicators.

[0090] In this embodiment, by using structured analysis and window screening to only retain stable and traceable data sources, fitting deviation caused by viscoelastic transition is avoided; and by using "turning interval" instead of single-point threshold to obtain effective hemostasis threshold (EHST), combined with the generation of safety deformation envelope (SDE) through spatiotemporal connectivity stability analysis, the efficacy and tissue safety are taken into account; within the uncertainty boundary, robust optimization is carried out, and individualized pressure-temperature setting with safety margin is directly output. Faster threshold reaching, lower concurrency (avoiding patching / penetration risk), higher robustness and repeatability can be achieved, and the cost of manual adjustment and trial and error is significantly reduced.

[0091] In an exemplary embodiment, according to the deformation field statistical data, the fitting parameter error range, and the stable data subset, the wound is subjected to spatiotemporal connectivity stability analysis to obtain safety deformation envelope data, including steps 502 to 512. Among them:

[0092] Step 502, the trend of the candidate stable period labeled by the stable data subset is analyzed jointly to determine the pressure maintaining segment stable window, which is flat, fluctuation convergent and spatially non-expanding.

[0093] Step 504, according to the deformation field statistical data and the fitting parameter error range, the strain threshold of the pressure maintaining segment stable window is subjected to adaptive segmentation analysis to obtain the strain mask sequence, the strain connected domain and the uncertainty envelope.

[0094] Step 506, according to the strain mask sequence and the uncertainty envelope, the wound is subjected to directional penetration / flow analysis to obtain penetration risk data and safety margin data.

[0095] Step 508, the counterfactual morphological stress test is performed on the high strain mask sequence and the uncertainty envelope to obtain robustness analysis data;

[0096] Step 510, according to the stable data subset and the strain mask sequence, the area, maximum span and occurrence time of the strain connected domain are subjected to time series analysis to obtain a hotspot residence expansion evidence map;

[0097] Step 512, according to the penetration risk data, the safety margin data, the robustness analysis data and the hotspot residence expansion evidence map, the wound is subjected to cascade threshold processing to obtain safety deformation envelope data.

[0098] Among them, the trend is flat, the circumferential average strain / maximum principal strain changes significantly in a short time window, and the rate of change with time is significantly reduced and maintained in a small range, which is manifested as a curve slope close to zero, which is used to determine that the deformation process enters the stable stage from the transition stage.

[0099] Wherein, the fluctuation convergence is that the variance / variation coefficient of the strain sequence in the same short window continuously decreases and has no sharp peak anomaly, indicating that the signal fluctuates around a stable mean value with small amplitude.

[0100] Wherein, the spatial non-expansion is that the number, area and maximum span of the connected domain of the high strain region in the short window do not continuously increase, and there is no new cluster or penetration, indicating that the risk hotspot does not expand.

[0101] Wherein, the steady-state window of the pressure maintaining section is a continuous time section that meets the conditions of "trend flattening, fluctuation convergence, and spatial non-expansion" under pressure maintaining conditions (and is guarded by pressure / visual).

[0102] Wherein, the adaptive segmentation analysis is to set the threshold and scale adaptively according to the fitting error and data distribution, and to process the strain field in different intensity intervals, so as to improve the robustness of segmentation and statistics.

[0103] Wherein, the strain mask sequence is a high strain region sequence obtained by binarizing each frame of strain field according to the adaptive threshold in the steady-state window.

[0104] Wherein, the strain connected domain is a connected high strain component labeled according to a preset neighborhood rule in a single frame of strain mask, which can be tracked in time sequence to count the area, span and appearance time length, etc.

[0105] Wherein, the uncertainty envelope is a pixel-level probability band (stable core / boundary band / exclusion band) formed by frequency statistics of multiple instance masks under perturbations such as threshold, calibration and reconstruction, which is used to quantify the confidence range of segmentation results.

[0106] Wherein, the directional penetration / seepage analysis is to examine whether the high strain mask forms a penetration path from the creation edge to the periphery under the constraint of the creation edge-periphery boundary, and to evaluate its forming conditions and path characteristics.

[0107] Wherein, the penetration risk data is the risk quantification result output by the directional seepage analysis, including whether it penetrates, the earliest triggering time, path length / width, etc., which is used for primary safety judgment.

[0108] Wherein, the safety margin data is the minimum fracture size and minimum shielding width required to break off the potential penetration in the case of no penetration, which reflects the remaining safety space from the risk boundary.

[0109] Wherein, the counterfactual morphological stress test is to implement multi-scale dilation / erosion and pose perturbation on the mask within the uncertainty range, and to observe the changes of the connectivity rate and cluster trend, in order to evaluate the robustness of the morphology under perturbation.

[0110] Wherein, the robustness analysis data is the quantitative score and label obtained by the counterfactual morphological stress test, which is used to represent the tendency of the hotspot to maintain or form a risk morphology under perturbation.

[0111] Among them, the hotspot residence expansion evidence map is an evidence map layer formed by accumulating the appearance length (residence) and area / span growth trend (expansion) in space after time tracking of the connected domain, used to identify "patchy and continuous" risk areas.

[0112] Among them, the cascade threshold processing is to sequentially threshold and fuse the through risk, safety margin, robustness score and residence expansion evidence according to priority, to remove unsafe areas and smooth the boundary, so as to generate the safety deformation envelope data.

[0113] Specifically, the candidate steady state period marked by the steady state data subset is aligned in phase with the pressure time series, the short window slope / variation rate evaluation (trend flattening) and short window variance / coefficient of variation evaluation (converging fluctuation) are performed on the circumferential average strain and maximum principal strain in the inherent coordinate of the creation edge, and the short window comparison (spatial non-expansion) is performed on the connected domain number, area and maximum span of the high strain area; each threshold is adaptively set according to the fitting parameter error range, and the pressure stability and visual quality guard (no sudden jump, no occlusion / reflection) are superimposed; the section that meets the three criteria and continuously maintains ≥ preset length is confirmed as the pressure maintaining segment steady state window.

[0114] In each pressure maintaining segment steady state window, first, the strain field is subjected to adaptive segmented thresholding according to the deformation field statistical data, to obtain the initial strain mask sequence; then, the threshold and displacement→strain reconstruction link are subjected to Monte Carlo / bootstrap disturbance (including threshold expansion, displacement field noise reconstruction, and calibration perturbation) according to the fitting parameter error range, to generate the target strain mask sequence; the pixel-level appearance frequency statistics of the target strain mask sequence are performed to form a confidence map, and the stable core, exclusion zone and boundary zone are obtained by double-threshold segmentation; the stable core is subjected to connected labeling to output the strain connected domain, and the boundary zone is encapsulated as an uncertainty envelope (including pixel-level probability and band width index) after morphological school accreditation (small scale inflation / erosion and bridge removal) along the creation edge-outer boundary direction.

[0115] Based on the strain mask sequence and the uncertainty envelope, a directed percolation graph is constructed under the constraint of the creation edge-outer boundary, to detect whether there is a high strain through path from the creation edge to the outer boundary for each pressure level / window; if there is, the through risk data (through labeling, earliest triggering time, path length / width) is output, and if there is not, the minimum fracture size and minimum shielding bandwidth required to break through the path are calculated as safety margin data.

[0116] Take the strain mask sequence as the benchmark, perform multi-scale dilation / erosion and random pose perturbation (translation / rotation / scale) on the mask within the amplitude allowed by the uncertainty envelope, and calculate the changes in statistical connectivity rate, coalescence rate, and path formation rate to form a robustness score for each connected domain and the global mask; the higher the score, the more likely it is to maintain or form risk patterns under perturbation. The output is the robustness analysis data and is backfilled into the corresponding pressure file / window.

[0117] According to the subset index of the steady-state data, the strain connected domains of the strain mask sequence are time-series tracked (associated according to the overlap / adjacency criterion), the time derivative (expansion trend) of the appearance length (residence), area and maximum span of each connected domain is calculated, and a residence and expansion evidence map aggregated by pixels / grids is generated; the evidence map is used to represent the "patchy and persistent" risk characteristics.

[0118] The through-risk data, safety margin data, robustness analysis data, and hotspot residence expansion evidence map are fused in the same pressure file, and the multi-layer fusion takes "whether through" as the first-level hard threshold, "safety margin ≥ threshold" and "robustness score ≤ threshold" as the second-level threshold, and "residence / expansion evidence below threshold" as the third-level threshold. The unsafe / edge region is excluded according to the rule from the inside to the outside; the through region is morphologically smoothed and overlaid with a conservative buffer zone, and the safety deformation envelope data (including spatial safety mask, risk mask, and recommended weight reduction / shielding partition) indexed by pressure file is output.

[0119] In this embodiment, the steady-state window is accurately locked by joint analysis of "trend flattening + fluctuation convergence + spatial non-expansion", avoiding the introduction of bias by transition period data; secondly, a more robust strain mask sequence and strain connected domain are obtained by adaptive segmentation and construction of uncertainty envelope; then, "whether through" is upgraded to a first-level hard criterion and the safety margin is quantified by directional through / pore flow analysis, replacing the traditional single-point threshold; combined with counterfactual morphological stress testing and hotspot residence expansion evidence, high-risk areas that are "patchy and persistent" are identified; finally, through cascade threshold fusion of multi-layer evidence, safety deformation envelope data that can be directly used for control is output. It can significantly reduce the misjudgment rate and through / overpressure risk, and improve the quantifiability, algorithm robustness and clinical repeatability of individualized safety boundaries.

[0120] In one exemplary embodiment, the wound is subjected to pressure-temperature cycle dynamic analysis based on the wound pressure-temperature optimization data to obtain wound pressure-temperature limit data, including steps 602 to 606. Among them:

[0121] Step 602, according to the wound pressure-temperature optimization data and the safety deformation envelope data, the pressure-temperature parameters of the wound rewarming cabin are analyzed in time domain channel to obtain pressure-temperature feasible channel data.

[0122] Step 604, according to the pressure-temperature feasible channel data and the fitting parameter error range, respectively, to the wound rewarming cabin of the pressure, constant pressure heating, maintenance period perturbation and pressure drop for the confrontation counterfactual scanning analysis, get the cycle boundary trajectory and trigger event set.

[0123] Step 606, the cycle boundary trajectory and trigger event set are processed by safety rhythm delimiting, and wound pressure-temperature limiting data is obtained.

[0124] Among them, the time domain channel analysis is to map the wound pressure-temperature optimization data (target pressure / temperature and partition weight, perturbation strategy) to the cycle template (pressure rise-constant pressure heating-maintenance-pressure drop) on the unified time axis, and intersect with the safety deformation envelope data at each time, get the pressure band, temperature band and boundary allowance available in each partition at each time / phase, at the same time, solidify the state guard and transition condition, form the analysis processing of the feasible search space for subsequent scanning.

[0125] Among them, the pressure-temperature feasible channel data is the result set of time domain channel analysis, recording "allowed pressure / temperature range, partition weight reduction / shielding mark, minimum safety allowance, state guard and transition condition" information according to time×partition; It converts "optimization setting + safety envelope" into time sequence feasible channel that can be directly traversed by algorithm.

[0126] Among them, the confrontation counterfactual scanning analysis is to construct the most unfavorable disturbance (threshold drift, calibrated perturbation, power deviation, measurement noise, etc.) in the search space of pressure-temperature feasible channel data, and gradually scan each sub-template of pressure rise, constant pressure heating, maintenance period perturbation and pressure drop, find the path and time that first touch the safety boundary (SDE) or trigger the risk mode.

[0127] Among them, the cycle boundary trajectory is the "nearest boundary point-time" sequence formed when the system state approaches the safety boundary along the feasible channel in the process of confrontation counterfactual scanning, which contains the corresponding pressure / temperature / partition combination, minimum safety allowance and normal crossing trend, which is used to quantify "how far away from risk" into executable limiting evidence.

[0128] Among them, the trigger event set is a discrete event set that is judged to reach or approach the risk threshold in scanning, such as SDE edge touch, through approaching, hot spot merging, rate overrun, temperature rise too fast, mass guard mismatch, etc.; Each event carries a timestamp, location / partition, trigger condition and associated boundary trajectory index.

[0129] The safety rhythm bounding process is a process of translating "cycle boundary trajectory + trigger event set" into executable timing constraints and guard rules, outputting the maximum / minimal ramp rate, minimum dwell time, allowed perturbation amplitude / frequency, partition weight reduction / shielding and prohibition combination list in each stage, and solidifying the automatic rollback / recovery strategy as a hard guardrail and scheduling contract for closed-loop control.

[0130] Specifically, the target pressure / temperature in the wound pressure temperature optimization data, the partition heating weight and the perturbation strategy are mapped to the cycle template (ramp-up, constant pressure temperature rise, maintain, ramp-down) of the unified time axis, and are intersected with the safety deformation envelope data at each time, forming the feasible interval or polygonal constraint set (allowed pressure band, allowed temperature band, partition weight reduction / shielding mark and boundary margin) of each partition at each time / phase, while recording the state guard (edge threshold, visual / sensing quality guard) and transition condition, and outputting the pressure temperature feasible channel data.

[0131] With the pressure temperature feasible channel data as the search space, the most unfavorable disturbance set (threshold drift, measurement noise, calibrated perturbation, partition power deviation, etc.) is constructed by combining the fitting parameter error range, and the four types of sub-templates of ramp-up, constant pressure temperature rise, maintenance period perturbation and ramp-down are scanned in each time step: the nearest point and normal crossing trend of the feasible region boundary are calculated in each time step, and the time stamp / position and corresponding boundary trajectory (including minimum safety margin and trigger path) of the first touch SDE boundary, penetration approach, hot spot and team, rate overrun, temperature rise too fast, etc. Trigger events are recorded to form the cycle boundary trajectory and trigger event set.

[0132] The cycle boundary trajectory and trigger event set are translated into executable constraints, including setting the maximum / minimal ramp rate and transition inhibition slope according to the earliest boundary point triggered in the ramp-up / down segment, setting the minimum dwell time and allowed perturbation amplitude / frequency according to the event recurrence interval and margin decay in the maintenance segment, and generating the prohibition list and partition weight reduction / shielding suggestion for the combination with frequent triggers; at the same time, the state guard (SDE edge, penetration signal, quality guard) and automatic rollback / recovery strategy are solidified, and finally the wound pressure temperature limiting data is outputted according to the stage x partition programming.

[0133] In this embodiment, by intersecting pressure-temperature optimization and safety deformation envelope at each time to obtain pressure-temperature feasible channel data, "what can be done" is determined at each time and partition; then, under the fitting uncertainty, the counterfactual scanning is carried out, the cycle boundary trajectory and trigger event set (touching edge, through approaching, merging, overheating, rate overrun, etc.) on the most unfavorable path are captured in advance, and the potential risk is converted into quantitative evidence; finally, the upper / lower limits of the ascending / descending rate, the minimum residence, the perturbation amplitude frequency, and the disabled combination of hard guardrails and supporting rollback / recovery strategies are generated by safety rhythm delimitation, which can significantly reduce the risk of boundary crossing and concurrency while ensuring efficacy, improve robustness, repeatability, and compliance, and reduce the cost of manual adjustment and trial and error.

[0134] In one exemplary embodiment, the wound tissue compliance data is nonlinearly fitted to obtain wound tissue fitting data, including steps 702 to 708. Among them:

[0135] Step 702, the hysteresis processing of the wound tissue compliance data is obtained to obtain a steady-state equivalent sample set.

[0136] Step 704, according to the steady-state equivalent sample set, the pressure-deformation relationship is nonlinearly fitted with shape prior and topological sparse constraint to obtain tissue preliminary fitting data.

[0137] Step 706, the tissue preliminary fitting data is solid boundary processed to obtain wound solid boundary fitting data.

[0138] Step 708, the wound solid boundary fitting data is structured and packaged to obtain wound tissue fitting data.

[0139] Among them, the hysteresis processing is a preprocessing procedure for stripping the loading / unloading hysteresis and viscoelastic tail effects of the pressure-deformation-time data collected on the ladder, which aligns the candidate steady-state period, closes the energy hysteresis, corrects the baseline drift, removes the jump / shading samples, and selects the representative points based on short window slope≈0 and variance convergence as criteria, so that the subsequent fitting is based on steady-state information without path dependence.

[0140] Among them, the steady-state equivalent sample set is the pressure and equivalent steady-state deformation pair and its confidence and time / step index extracted by the step pressure holding window after hysteresis processing.

[0141] Among them, the pressure-deformation relationship is the function / segmented mapping of the surface deformation (or strain) response of the tissue under the action of the applied pressure under the inherent coordinates of the wound edge and the unified time axis, which is used to describe the mechanical compliance behavior of "easy compression → difficult compression", and is the basis object for threshold identification and control setting.

[0142] Wherein, the shape prior is a physical consistency constraint imposed on the pressure-deformation relationship, including monotonic non-decreasing, amplitude limiting (upper and lower bounds), fast first and slow later (slope decreasing), and smooth continuity, etc., to avoid over-fitting curves without physical meaning and stabilize the turning point identification.

[0143] Wherein, the topological sparse constraint is a regularization strategy (minimum turning segment / minimum topological complexity) to minimize the number and complexity of segments / turning points when modeling the curve, which, while satisfying the shape prior, suppresses false turning points caused by noise, making the model concise and interpretable.

[0144] Wherein, the preliminary fitting data is the initial nonlinear fitting data obtained based on the steady-state equivalent sample set under the shape prior and topological sparse constraint, including continuous and derivable curve expression and its error / residual information and candidate turning point hints.

[0145] Wherein, the boundary handling is a checking process of bounding projection and worst-case disturbance back-projection on the preliminary fitting results under the constraints of device capability and safety boundaries (such as safety deformation envelope, maximum allowed slope / curvature, etc.), which prohibits extrapolation beyond the boundary, forces monotonic amplitude limiting, and calculates the remaining margin to the safety boundary.

[0146] Wherein, the wound boundary fitting data is the fitting result set that satisfies both physical and safety constraints after boundary handling, with compatibility labels and minimum boundary margins, which can be directly used for threshold identification, optimization and control without further safety checking.

[0147] Wherein, the structured packaging process is to organize the wound boundary fitting data and its metadata (source steps, steady-state evidence references, residual summaries, version timestamps, etc.) into standardized, callable interfaces and data formats (such as query pressure-deformation / slope, inverse pressure, turning zone query, margin query), so that subsequent double-threshold analysis, channel construction and closed-loop control modules can be directly consumed.

[0148] Specifically, in the original pressure-deformation-time sequence of the wound tissue compliance data representation ladder, first align the candidate steady-state period and pressure guard, identify the loading / unloading hysteresis and viscoelastic tailing zone; perform energy hysteresis closure and baseline drift correction on the transition segment, eliminate abnormal points such as sudden jumps / occlusions, and select representative points for each step of pressure preservation based on the criteria of short window slope ≈ 0 and variance convergence; finally, the pressure, equivalent steady-state deformation and its confidence, timestamp and source index of each step are aggregated to form a steady-state equivalent sample set that is free of path dependence.

[0149] Under monotonic non-decreasing, clipping (upper / lower bound) and "fast first and slow later" shape priors, piecewise smooth fitting (such as restricted spline / segmented exponential or Gompertz / generalized Logistic family) is performed on the steady-state equivalent sample set, and the "minimum turning section" is used as the sparse regularity to control the complexity, and the number of segments and the smoothing strength are automatically selected by cross-validation; the continuous and derivable fitting curve and the segment boundary position, residual spectrum and candidate turning point are output, which constitute the preliminary fitting data of the organization.

[0150] Under the constraints of device capability and organizational safety boundary, the preliminary fitting data of the organization is subjected to bounding projection, limiting the maximum slope and curvature (Lipschitz and second-order upper bound), prohibiting extrapolation beyond the boundary, forcing monotonicity and non-negativity, and projecting the curve back to the feasible region under the most unfavorable disturbance scenario (threshold drift, calibration perturbation); the minimum boundary margin of the safety deformation envelope / device upper limit is calculated and labeled with a compatibility tag to obtain wound boundary fitting data that meets both physical and safety constraints.

[0151] The wound boundary fitting data after boundarying and its metadata are packaged into a callable interface, providing functions such as "given pressure to find deformation / slope / curvature, given deformation to find pressure, turning interval query, boundary margin query"; at the same time, audit information such as source step index, steady-state evidence reference, residual summary, version and timestamp is encapsulated and serialized into a standard format to form wound tissue fitting data that can be directly consumed by double-threshold joint analysis, time-domain channel and subsequent control.

[0152] In this embodiment, the steady-state equivalent samples without transition tail are first generated by the de-lag processing, which significantly reduces the fitting deviation; the nonlinear fitting under the shape prior and topological sparse constraint obtains a simple, interpretable and abnormality-insensitive individualized pressure-deformation relationship; the boundarying process projects the most unfavorable disturbance, device limit and safety deformation boundary onto the model, and outputs the "executable curve" with boundary margin, which suppresses the risk of overpressure and boundary crossing from the source; finally, the structured packaging is encapsulated into a standard interface, which facilitates the direct call of subsequent double-threshold analysis, time-domain channel and closed-loop control, and realizes higher precision and robustness, lower parameter tuning cost and stronger repeatability and traceability.

[0153] In one exemplary embodiment, according to the continuous imaging data of the wound edge safety area, the surface displacement strain time series field of the wound is calculated, including steps 802 to 808. Among them:

[0154] Step 802, the continuous imaging data of the wound edge safety area and the baseline imaging data are subjected to medium geometric reduction processing to obtain a sequence of wound edge intrinsic coordinate frames.

[0155] Step 804, the double-flow counterfactual reduction processing is performed on the sequence of wound edge intrinsic coordinate frames to obtain a sub-pixel displacement time series field.

[0156] Step 806, topological physical reduction processing is performed on the sub-pixel displacement time sequence field to obtain an initial surface strain time sequence field.

[0157] Step 808, quality control reduction processing is performed on the initial surface strain time sequence field to obtain a surface displacement strain time sequence field.

[0158] Wherein, the medium geometry reduction processing is to complete the distortion correction, refraction correction and parallax compensation under the constraints of the pose calibration result and the transparent airbag optical model (camera internal and external parameters, lens distortion, cross-medium refraction / curvature), and to construct the inherent coordinate (such as polar coordinate / flattened grid) and unified ROI with the reference of the wound edge boundary, so that all subsequent images fall on the same, geometric-optical consistent analysis plane.

[0159] Wherein, the wound edge inherent coordinate frame sequence is a time sequence image set formed by mapping the baseline / continuous frame to the inherent coordinate system with the wound edge as the origin or reference after the medium geometry reduction processing, which already contains the boundary mask, pixel-physical scale mapping and time alignment information.

[0160] Wherein, the double-flow counterfactual reduction processing is to calculate the related matching (DIC) and optical flow two-way displacement candidates on the wound edge inherent coordinate frame sequence in parallel, to strip and separate the motion components irrelevant to pressure from the "pressure-free counterfactual" trajectory generated by the baseline frame, and to dynamically weight and sub-pixel refine the double-flow with consistency residual gating, so as to only retain the true displacement components induced by pressure.

[0161] Wherein, the sub-pixel displacement time sequence field is a displacement vector field sequence obtained after the double-flow counterfactual reduction processing, which is continuous in time axis and sub-pixel accuracy in space, and is accompanied by residual / error and confidence information of each frame (or each pixel).

[0162] Wherein, the topological physical reduction processing is to introduce boundary conservation, radial compliance monotony, shear smoothing and topological preservation (non-splitting / non-pseudo-merging) physical-topological consistency constraints when converting the sub-pixel displacement field to the strain field, and to inhibit noise and non-physical gradients through regularization projection, so as to ensure that the reconstructed strain is consistent with the anatomical boundary and actual stress.

[0163] Wherein, the initial surface strain time sequence field is a surface strain field sequence obtained after the topological physical reduction processing, which is consistent with the wound edge anatomical boundary and continuous in time and space.

[0164] Wherein, the quality control reduction processing is to perform low-confidence frame rejection, physically consistent time sequence interpolation, color scale / dimension standardization and time stamp registration on the initial strain time sequence according to the residual and confidence threshold, baseline alignment and visual quality safeguard (obstruction / reflection / fogging / film grain interference), and finally output the high-confidence "surface displacement-strain time sequence field" that can be used for criteria and control.

[0165] Specifically, based on the pose calibration results and the optical geometric model of the transparent bladder (camera intrinsic and extrinsic parameters, lens distortion, cross-medium refraction / curvature), the continuous imaging data of the safety area of the wound edge and the baseline imaging data are first corrected frame by frame for distortion, refraction, and parallax compensation; then the wound edge boundary is taken as a reference to construct an intrinsic coordinate (such as polar coordinates / flattened grid), and the two types of images are uniformly mapped to the same wound edge intrinsic coordinate system and time synchronization and ROI mask solidification are completed, and the wound edge intrinsic coordinate frame sequence that is geometric-optically consistent is output.

[0166] On the wound edge intrinsic coordinate frame sequence, the digital image correlation (DIC) and optical flow two-way displacement candidates are calculated in parallel for adjacent frames (and with the baseline frame), and a double-flow consistency residual map is generated; the baseline imaging data is used to deduce the no-pressure counterfactual trajectory, and the background / respiration components unrelated to pressure are stripped, and only the pressure-induced motion is retained; then the DIC / optical flow is dynamically weighted and fused and sub-pixel level sub-window refined with residual gating, to suppress single-path deviation and noise drift, and form a physically consistent and time-continuous sub-pixel displacement time sequence field.

[0167] The sub-pixel displacement time sequence field is converted into strain through a topological-physically consistent strain reconstruction link, that is, under the prior conditions of wound edge boundary conservation, radial monotonicity and shear smoothness, finite strain / small strain calculation is adopted combined with topological preservation (connectedness / non-cracking) and regularization projection (such as TV / Laplace constraint) to suppress crack-like artifacts and cluster distortion; then the abnormal gradient is robustly suppressed and locally projected, and an initial surface strain time sequence field consistent with the anatomical boundary and spatially continuous is output.

[0168] According to the previously retained residual / confidence information and baseline alignment threshold, the initial surface strain time sequence field is quality controlled, that is, low-confidence frames (obstruction, reflection, fogging, film grain interference) are marked and removed, and short frames are repaired by physically consistent time interpolation; then the whole sequence is dimensionally / color-scaled standardized and pressure time stamp aligned to ensure that different stages can be compared horizontally; finally, the surface displacement strain time sequence field that passes quality control (for threshold determination, compliance identification, and safety envelope analysis) is obtained, with frame-level quality labels attached.

[0169] In this embodiment, the refraction / distortion of the transparent balloon is eliminated by medium geometric reduction, and the system error introduced by the carrier and the viewing angle is avoided; then the non-pressurized component is stripped by related matching and double-flow fusion combined with baseline counterfactual, to obtain sub-pixel accuracy, time-continuous causal displacement; then the strain is reconstructed under the constraints of boundary conservation, radial compliance monotony and topological preservation, to suppress non-physical artifacts such as cracking and clustering, and to ensure consistency with anatomy and stress; finally, the surface displacement-strain time series field which can be directly used for threshold judgment, compliance fitting and safety envelope generation is output by completing rejection and interpolation, dimension and time alignment, under the guardianship of residual / error / confidence and visual quality. The measurement accuracy, robustness and repeatability can be significantly improved, and the risk of clinical adjustment and misjudgment can be reduced.

[0170] In an exemplary embodiment, the sequence of wound margin intrinsic coordinate frames is processed by double-flow counterfactual reduction to obtain a sub-pixel displacement time series, including steps 902 to 910. Among them:

[0171] Step 902, the sequence of wound margin intrinsic coordinate frames and baseline imaging data are processed by homophase, to obtain a sequence of time frame pairs and a wound margin time index;

[0172] Step 904, the relevant features and optical flow features in the sequence of time frame pairs are double-flow refined to obtain a consistent residual map, a relevant displacement flow and an optical flow displacement flow;

[0173] Step 906, according to the wound margin time index and the consistent residual map, the relevant displacement flow and the optical flow displacement flow are processed by causal stripping to obtain a pressurization-induced displacement time series;

[0174] Step 908, the pressurization-induced displacement time series and the consistent residual map are processed by cooperative distillation to obtain a sub-pixel displacement candidate field;

[0175] Step 910, according to the wound margin boundary priori constraint and the radial compliance constraint, the sub-pixel displacement candidate field is processed by constraint projection to obtain a sub-pixel displacement time series field.

[0176] The homophase processing is a step of fine-grained alignment of the sequence of wound margin intrinsic coordinate frames and the baseline imaging according to time stamp, pressure record and phase correlation, which eliminates the acquisition rhythm difference and trigger jitter by resampling and phase correction, so that any observation frame at any time can be "one-to-one paired" with the corresponding baseline frame.

[0177] The sequence of time frame pairs is a paired image set arranged in time sequence after homophase, each pair is composed of "current frame-corresponding baseline or adjacent frame", and carries unified time stamp and pressure stage information.

[0178] Wherein, the creation edge timing index is an index entry established for each time frame pair, recording its belonging pressure stage, step window, frame number interval and ROI mask, etc. metadata, facilitating quick backtracking and step-by-step reference in causal stripping and event positioning.

[0179] Wherein, the relevant feature is a texture block / descriptor and its matching response (such as cross-correlation peak, confidence, sub-window interpolation displacement) for digital image correlation (DIC), used to measure the local block correspondence and displacement estimation between frame pairs.

[0180] Wherein, the optical flow feature is the pixel-level motion information (such as pyramid optical flow vector, gradient intensity, residual and visibility label) calculated based on brightness consistency and gradient constraint, used to depict the fine-grained displacement field between consecutive frames.

[0181] Wherein, the dual-flow extraction is to extract "relevant features" and "optical flow features" for each time frame pair simultaneously, respectively solving two displacement estimates, and combining local texture / confidence to generate consistency metrics.

[0182] Wherein, the relevant displacement flow is a displacement vector field obtained by block matching and sub-pixel interpolation from relevant features, good at handling areas with clear texture, and accompanied by matching confidence and occlusion hints.

[0183] Wherein, the optical flow displacement flow is a displacement vector field obtained by pyramid multi-scale solving from optical flow features, more sensitive to small continuous motion and weak texture areas, and accompanied by residual and stability indicators.

[0184] Wherein, the consistency residual map is a difference intensity map (including threshold / confidence information) obtained by pixel-by-pixel comparison of relevant displacement flow and optical flow displacement flow, used to mark areas where the two estimates are inconsistent or affected by occlusion / reflection, guiding the weight allocation of subsequent stripping and fusion.

[0185] Wherein, the causal stripping process is to remove respiratory, body motion, camera drift, etc. non-stress components from the two displacement flows, only retaining the causal motion components caused by pressure changes, with the pressure stage annotated by the creation edge timing index as the time reference, and the "no-pressure prediction trajectory" generated from the baseline frame.

[0186] Wherein, the stress-induced displacement timing is a sequence of displacement vector fields arranged over time after causal stripping, representing "only pressure-induced" true tissue displacement.

[0187] Wherein, the collaborative distillation process is to use the consistency residual map for gating, and to fuse and refine the relevant displacement flow and the optical flow displacement flow with spatial adaptive weights and temporal filtering (such as smoothing / Kalman update), to suppress single-path bias, forming a candidate displacement with lower noise and better continuity.

[0188] The sub-pixel displacement candidate field is a high-precision, time-space continuous displacement vector candidate result output by the cooperative distillation process, and interpolation and smoothing have been completed at the sub-pixel level.

[0189] The wound edge boundary prior constraint is a normal / tangential behavior constraint (such as boundary normal continuity, no outward turning and tangential smoothness) established according to the wound edge anatomical boundary and imaging geometry, and is used to ensure the physical consistency of the displacement field at the boundary.

[0190] The radial compliance constraint is to decompose the displacement into radial / tangential components with the wound edge as a reference, and to impose monotonicity, amplitude limiting and priority requirements on the radial component to meet the mechanical properties of the compressed tissue “complying inward”.

[0191] The constraint projection process is to project the sub-pixel displacement candidate field to the feasible manifold that satisfies the wound edge boundary prior and radial compliance constraint, and to make the minimum change amount back to the vector that violates the constraint until the physical-topological consistency of the whole field is achieved.

[0192] Specifically, the wound edge intrinsic coordinate frame sequence and the baseline imaging data are synchronously resampled according to the time stamp and pressure record, and the phase correlation / cross-correlation peak is used to perform fine-grained phase correction on the observation at each time and its baseline reference; After alignment, a one-to-one corresponding frame pair sequence is generated as a time frame pair sequence, while outputting the wound edge time sequence index recording “pressure stage-time window-frame number”.

[0193] For each pair of aligned frames in the time frame pair sequence, multi-scale digital image correlation (DIC) and pyramid optical flow estimation are performed respectively, and sub-window interpolation is used to refine the sub-pixel displacement; The consistency residual map is calculated by subtracting the two displacement at the pixel level and combining the local texture confidence, while retaining the confidence weight and occlusion prompt of each other, forming the correlation displacement flow and the optical flow displacement flow.

[0194] According to the wound edge time sequence index, determine the pressure stage where each frame is located, and use the non-pressure prediction trajectory generated by the baseline reference frame to strip the non-pressurized components such as background and respiration from the correlation displacement flow and the optical flow displacement flow; Then gated by the consistency residual map, shield the low consistency area and constrain the residual displacement in the time direction monotonicity and the physical direction (radial priority), output the pressure-induced displacement time sequence that only retains “due to pressure”.

[0195] The information in the pressure-induced displacement time sequence and the consistency residual map is dynamically weighted according to the spatial adaptive weight; Cooperate with the exponential smoothing and Kalman type update in time to suppress noise drift, and use spline / Lagrange interpolation to refine at the sub-pixel level, to get the time-space continuous, locally smooth sub-pixel displacement candidate field.

[0196] According to the boundary prior constraint and the radial compliance constraint, the sub-pixel displacement candidate field is projected to the physical feasible manifold. In the projection process, the radial component is forced to be monotonic and non-decreasing, the tangential component is applied with smoothing / divergence suppression, the normal continuity and non-outward turning constraints are performed at the boundary, and the vectors that violate the constraints are projected back with the minimum change. The sub-pixel displacement time sequence field that satisfies the boundary conservation and the radial compliance is output after iteration frame by frame to convergence.

[0197] In the embodiment, by eliminating the acquisition rhythm difference and trigger jitter through homodyning, false displacement caused by time mismatch is avoided; by using double-flow extraction, the correlation / DIC and optical flow are complemented, and the unreliable area is marked by a consistent residual map, the lower limit of displacement estimation and interpretability are improved; further, using causal stripping, the respiratory / body motion and other non-pressurization components are removed under the boundary time sequence index constraint, only the pressurization-induced motion is retained, and the systematic interference is significantly reduced; and in cooperation with distillation, adaptive weighting and time filtering are performed using residual gating to obtain sub-pixel candidates with lower noise and better continuity; finally, under the boundary prior and radial compliance constraints, the candidates are projected to the physically feasible manifold to suppress artifacts such as cracking and coalescence and ensure boundary conservation. The accuracy, robustness and repeatability are significantly improved, and an auditable and directly callable causal displacement foundation is provided for subsequent strain reconstruction and safety control.

[0198] In one exemplary embodiment, a deformation visualization monitoring system for wound rewarming cabin pressurization process is provided, the system comprising a terminal and a computer device;

[0199] The main core functions of the computer device include:

[0200] 1. Heat preservation module: using ultra-thin flexible graphene material, 3 seconds of rapid heating, uniform heat coverage of the wound surface, avoiding local burns or uneven temperature caused by traditional hot compress, even if the external temperature drops suddenly, the cabin can still maintain stable temperature for 6-12 hours. The outer layer is a space-level aerogel thermal insulation layer, combined with vacuum sealing technology, effectively isolating external cold air, ensuring continuous heat preservation for 6-12 hours in extremely cold environments.

[0201] 2. Control module: improved electronic sphygmomanometer control panel design, with lithium battery (can standby for 8 hours at -30℃) integrating Bluetooth and mobile phone App into the control chip (memory system can be implanted on the chip, which can efficiently heat up and pressurize to stop bleeding) Double insurance can also be used to manually control the interface for emergency. The charging source can be charged through the UTC port, and if in the field in extremely cold environments, it can be charged by solar energy. A connection port is arranged at the bottom end of the control panel (a spare device is arranged at the outer end), and a modified wrist sphygmomanometer is connected through a data line to realize data intercommunication.

[0202] 3. Heating / pressurizing module: Add TPE high-transparency, flexible airbags and mesh heating wires outside the insulation module. Precise control of airbag pressure 30-40 mmHg and temperature 36-38℃ through control board.

[0203] 4. Fixing module: It is recommended to use a light, heat-insulating, waterproof, wear-resistant, and portable Kordura nylon + aerogel composite material. High-strength, low-temperature-resistant, and single-handed quick-release buckle-type fixation is recommended for both ends. The outer layer can be opened with a 10X10cm size window, which is sealed at the edge for easy observation of the wound condition.

[0204] 5. Life monitoring module: It uses high-strength cold-resistant composite webbing, with an outer layer of wear-resistant and tear-resistant tactical nylon material and an inner layer of embedded flexible silicone anti-slip layer. When worn on the wrist, it is both stable and not harsh on the skin. The webbing is equipped with an adaptive adjustment buckle made of cold-resistant rubber, which remains flexible in low-temperature environments. In extreme environments, it reliably secures the device, ensuring continuous and stable contact between the monitoring module and the skin, and ensuring accurate collection of vital signs data. Heart rate and blood oxygen saturation monitoring use advanced photoplethysmography (PPG) technology based on blood absorption changes to ensure that heart rate and blood oxygen saturation monitoring data are accurately controlled within ±5% and ±2%, respectively. Blood pressure monitoring uses PPG (green LED + photodiode) to monitor blood pressure volume changes through optical principles, combined with pressure sensor data, with a measurement error of ±8mmHg. Respiratory rate can be measured by combining PPG green LED irradiation with electrocardiogram (ECG) sensors, which capture the subtle fluctuations in blood vessel volume with respiration and simultaneously collect electrocardiogram signals to accurately calculate respiratory rate. Body temperature monitoring uses an infrared thermopile sensor, which can measure temperature non-contact, avoiding the risk of metal probes sticking to the skin in low-temperature environments.

[0205] In an example embodiment, a computer device, which can be a server, has an internal structure diagram as shown in Figure 3 The computer device includes a processor, a memory, an input / output interface (I / O), and a communication interface. Those skilled in the art can understand that the structure shown in Figure 3 is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. A specific computer device can include more or fewer components than those shown in the diagram, or combine certain components, or have a different arrangement of components.

[0206] In an embodiment, a computer device is also provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps in the above method embodiments.

[0207] In one embodiment, a computer readable storage medium storing a computer program is provided, the computer program, when executed by a processor, implements the steps in the above method embodiments.

[0208] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steps in the above method embodiments.

[0209] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0210] A person of ordinary skill in the art can understand that all or part of the processes in the above method embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above method embodiments.

[0211] The technical features of the above embodiments can be combined in any manner. To make the description concise, all possible combinations of the technical features in the above embodiments are not described, but as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present application.

[0212] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for visually monitoring deformation during the pressurization process of a wound rewarming chamber, characterized in that, The method includes: With the wound rewarming chamber open, baseline imaging data of the safe area of ​​the wound edge is obtained based on the optical geometry of the pose calibration camera and the transparent airbag. The wound rewarming chamber is controlled to apply test pressure to the wound, and the surface displacement strain time series field of the wound is calculated based on the continuous imaging data of the wound edge safety area. Based on the surface displacement strain time sequence field, the wound rewarming chamber is controlled to apply stepwise pressure to the wound to obtain wound tissue compliance data. Based on the wound tissue compliance data, the wound rewarming chamber is controlled to perform pressure-temperature coupled circulation on the wound to obtain wound tissue maintenance data. Based on the wound tissue maintenance data, pressure-temperature coupling perturbation correction is performed on the wound to obtain wound tissue stability data; the wound tissue stability data is used to generate deformation visualization data of the wound rewarming chamber.

2. The method according to claim 1, characterized in that, The step of controlling the wound rewarming chamber to perform pressure-temperature coupled circulation on the wound based on the wound tissue compliance data to obtain wound tissue maintenance data includes: Nonlinear fitting is performed on the wound tissue compliance data to obtain wound tissue fitting data; Based on the wound tissue fitting data, a dual-threshold joint analysis of the wound treatment was performed to obtain wound pressure and temperature optimization data; Using the wound pressure and temperature optimization data as a reference, a dynamic analysis of the wound pressure and temperature cycle was performed to obtain the wound pressure and temperature limit data. Based on the wound pressure and temperature optimization data and the wound pressure and temperature limitation data, the wound rewarming chamber is controlled to perform pressure and temperature coupling cycles on the wound to obtain the wound tissue maintenance data.

3. The method according to claim 2, characterized in that, The step of performing a dual-threshold joint analysis on the wound based on the wound tissue fitting data to obtain wound pressure-temperature optimization data includes: The wound tissue fitting data were subjected to structured analysis to obtain compliance curve data, deformation field statistics, fitting parameter error range, and pressure holding steady-state evidence. The steady-state evidence under pressure holding is filtered by a window to obtain a subset of steady-state data; Based on the compliance curve data and the error range of the fitting parameters, the turning interval of the wound is analyzed to obtain the effective hemostatic threshold pressure data. Based on the deformation field statistics, the fitting parameter error range, and the steady-state data subset, a spatiotemporal connectivity stability analysis is performed on the wound to obtain safe deformation envelope data. The effective hemostasis threshold pressure data and the safe deformation envelope data are robustly optimized to obtain the wound pressure-temperature optimized data.

4. The method according to claim 3, characterized in that, The step of performing spatiotemporal connectivity stability analysis on the wound based on the deformation field statistics, the fitting parameter error range, and the steady-state data subset to obtain safe deformation envelope data includes: A joint analysis of the trend flattening, fluctuation convergence, and spatial non-expansion of the candidate steady-state periods labeled in the aforementioned steady-state data subset is performed to determine the steady-state window of the pressure-holding segment; Based on the deformation field statistics and the error range of the fitting parameters, an adaptive segmentation analysis is performed on the strain threshold of the steady-state window of the pressure holding section to obtain the strain mask sequence, strain connected domain, and uncertainty envelope. Based on the strain mask sequence and the uncertainty envelope, a directional penetration / seepage analysis is performed on the wound to obtain penetration risk data and safety margin data. Counterfactual morphological stress tests were performed on the strain mask sequence and the uncertainty envelope to obtain robustness analysis data. Based on the steady-state data subset and the strain mask sequence, time series analysis is performed on the area, maximum span, and duration of occurrence of the strain connectivity domain to obtain a hotspot dwelling expansion evidence map. Based on the penetration risk data, the safety margin data, the robustness analysis data, and the hotspot residence and expansion evidence map, the wound is subjected to cascaded threshold processing to obtain the safety deformation envelope data.

5. The method according to claim 3, characterized in that, The step of performing dynamic analysis of wound pressure and temperature cycles on the wound, using the optimized wound pressure and temperature data as a reference, to obtain wound pressure and temperature limit data includes: Based on the wound pressure-temperature optimization data and the safety deformation envelope data, time-domain channel analysis is performed on the pressure-temperature parameters of the wound rewarming chamber to obtain feasible pressure-temperature channel data. Based on the feasible pressure-temperature channel data and the error range of the fitting parameters, adversarial counterfactual scanning analysis was performed on the pressurization, constant pressure heating, maintenance period perturbation and depressurization of the wound rewarming chamber to obtain the cycle boundary trajectory and trigger event set. The loop boundary trajectory and the trigger event set are subjected to safety rhythm delimitation processing to obtain the wound pressure and temperature limit data.

6. The method according to claim 2, characterized in that, The step of performing nonlinear fitting on the wound tissue compliance data to obtain wound tissue fitting data includes: The wound tissue compliance data were dehysterized to obtain a steady-state equivalent sample set; Based on the steady-state equivalent sample set, the pressure-deformation relationship is subjected to nonlinear fitting processing with shape prior and topological sparsity constraints to obtain preliminary tissue fitting data. The preliminary tissue fitting data is subjected to fixation processing to obtain wound fixation fitting data. The wound solid boundary fitting data is structured and encapsulated to obtain the wound tissue fitting data.

7. The method according to claim 1, characterized in that, The step of calculating the surface displacement strain time-series field of the wound based on continuous imaging data of the wound edge safety area includes: The continuous imaging data of the wound edge safety area and the baseline imaging data are subjected to medium geometry reduction processing to obtain the wound edge intrinsic coordinate frame sequence. The original coordinate frame sequence is subjected to dual-stream antifactual reduction to obtain the subpixel displacement temporal field; The subpixel displacement temporal field is subjected to topological physical reduction to obtain the initial surface strain temporal field; The initial surface strain time series field is subjected to quality control reduction processing to obtain the surface displacement strain time series field.

8. The method according to claim 7, characterized in that, The process of performing dual-stream counterfactual reduction on the intrinsic coordinate frame sequence to obtain the sub-pixel displacement temporal field includes: The inherent coordinate frame sequence of the wound edge and the baseline imaging data are subjected to in-phase processing to obtain a time frame pair sequence and a wound edge temporal index. The correlation features and optical flow features in the time frame sequence are extracted using a dual-stream method to obtain the correlation displacement flow, optical flow displacement flow, and consistency residual map. Based on the creation timing index and the consistency residual map, the causal stripping process is performed on the relevant displacement flow and the optical displacement flow to obtain the pressure-induced displacement timing. The pressure-induced displacement time series and the consistency residual map are subjected to co-distillation processing to obtain a sub-pixel displacement candidate field; Based on the prior constraints of the creation boundary and the radial compliance constraints, the sub-pixel displacement candidate field is subjected to constraint projection processing to obtain the sub-pixel displacement temporal field.

9. A deformation visualization monitoring system for the pressurization process of a wound rewarming chamber, characterized in that, The system includes: a terminal and computer equipment; The computer device is used to acquire baseline imaging data of the safe area of ​​the wound edge based on the optical geometric relationship between the pose calibration camera and the transparent airbag when the wound rewarming chamber is not closed; the terminal is used to detect the closure status of the wound rewarming chamber. The computer device is used to control the wound rewarming chamber to perform test pressurization on the wound, and to calculate the surface displacement strain time-series field of the wound based on continuous imaging data of the wound edge safety area. The computer device is used to control the wound rewarming chamber to apply stepwise pressure to the wound based on the surface displacement strain time sequence field, so as to obtain wound tissue compliance data. The computer device is used to control the wound rewarming chamber to perform pressure-temperature coupling circulation on the wound based on the wound tissue compliance data, so as to obtain wound tissue maintenance data. The computer device is used to perform pressure-temperature coupling perturbation correction on the wound based on the wound tissue maintenance data to obtain wound tissue stability data; the wound tissue stability data is used to generate deformation visualization data of the wound rewarming chamber.

10. The system according to claim 9, wherein the computer device includes a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images