Deformation visual monitoring method and system in pressurization process of wound rewarming cabin

Abstract

The invention relates to a deformation visual monitoring method and system in the pressurization process of a wound rewarming cabin. The method comprises the following steps: acquiring baseline imaging data of a wound margin safety area according to an optical geometrical relationship between a pose calibration camera and a transparent airbag; controlling the wound rewarming cabin to perform testability pressurization on the wound, and calculating a surface displacement strain time sequence field of the wound according to the continuous imaging data of the wound margin safety area; according to the surface displacement strain time sequence field, the wound rewarming cabin is controlled to conduct stepped upward pressing on the wound, and wound tissue compliance data are obtained; according to the wound tissue compliance data, the wound rewarming cabin is controlled to conduct pressure-temperature coupling circulation on the wound, and wound tissue maintainability data is obtained; according to the wound tissue maintainability data, performing pressure-temperature coupling perturbation correction on the wound to obtain wound tissue stability data; the method is used for generating deformation visualization data of the wound rewarming cabin. By adopting the method, the treatment success rate of the wounded in the cold region can be effectively improved.

Description

Technical Field

[0001] This application relates to the field of automatic control technology, and in particular to a method and system for visually monitoring deformation during the pressurization process of a wound rewarming chamber. Background Technology

[0002] During activities in cold regions, the environment is extremely harsh, with frequent occurrences of extreme conditions such as low temperatures, strong winds, and blizzards. After injury, the body is highly susceptible to hypothermia due to low temperatures, leading to slow wound healing, abnormal blood clotting, and various complications, significantly increasing the difficulty of treatment and mortality. Traditional vital sign monitoring devices, such as ordinary life monitoring bracelets, not only suffer from significantly reduced accuracy in low-temperature environments but also have limited functionality, making it difficult to comprehensively and in real-time monitor the vital signs of the injured, thus failing to provide timely and effective data support for treatment. Furthermore, existing wound rewarming devices often operate independently, lacking effective linkage with the vital sign monitoring system, and cannot adjust rewarming strategies in a timely manner based on dynamic changes in the injured's vital signs, resulting in a low success rate in treating injured individuals during activities in cold regions. Summary of the Invention

[0003] Therefore, it is necessary to provide a method and system for visually monitoring deformation during the pressurization process of a wound rewarming chamber, which can effectively improve the success rate of treating injured people in cold regions.

[0004] Firstly, this application provides a method for visually monitoring deformation during the pressurization process of a wound rewarming chamber, including: 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.

[0005] Secondly, this application also provides a deformation visualization monitoring system for the pressurization process of a wound rewarming chamber, the system including 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.

[0006] The aforementioned method and system for visually monitoring deformation during the pressurization process of a wound rewarming chamber first establishes the optical geometric relationship between the camera and the transparent airbag using pose calibration while the wound rewarming chamber is not closed, acquiring baseline imaging data of the safe area at the wound edge to ensure non-invasiveness and safety during the imaging process. Then, under controlled conditions, the wound is subjected to test pressurization, and the surface displacement strain time-series field is calculated based on continuous imaging results, enabling the dynamic and precise capture of minute deformations of the wound tissue. Further, a stepped pressurization test is conducted to obtain tissue compliance data, providing a scientific basis for evaluating the biomechanical response of the wound under different external forces. Subsequently, a pressure-temperature coupled cyclic test is combined to obtain tissue maintenance data, reflecting the functional recovery capacity of the wound tissue under comprehensive stress conditions from multiple dimensions. Finally, through pressure-temperature coupled perturbation correction, data visualization results that intuitively demonstrate wound stability are generated. It not only enables the systematic and quantitative assessment of key parameters such as wound tissue strain characteristics, compliance, and stability, but also provides a scientific and reliable basis for the development of personalized treatment strategies in clinical practice, greatly improving the safety, specificity, and controllability of wound treatment, and effectively increasing the success rate of treatment for injured patients in cold regions. Attached Figure Description

[0007] 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.

[0008] 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; Figure 2 This is a flowchart illustrating a deformation visualization monitoring method for the pressurization process of a wound rewarming chamber in one embodiment. Figure 3 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0009] 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.

[0010] 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.

[0011] 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:

[0012] 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.

[0013] 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.

[0014] Step 206: Based on the surface displacement strain time sequence field, control the wound rewarming chamber to apply stepwise pressure to the wound to obtain wound tissue compliance data.

[0015] Step 208: Based on the wound tissue compliance data, control the wound rewarming chamber to perform pressure-temperature coupling circulation on the wound to obtain wound tissue maintenance data.

[0016] Step 210: Based on the wound tissue maintenance data, perform pressure-temperature coupling perturbation correction on the wound to obtain wound tissue stability data.

[0017] The wound rewarming chamber is a closed device used to implement pressure and temperature control treatment on the wound. It includes a transparent airbag (pressurization medium), a zoned heating / temperature control module, pressure and temperature sensors, an imaging window and a camera bracket, etc. Its core function is to apply a programmable pressure-temperature trajectory to the wound according to the control strategy.

[0018] Among them, the pose calibration camera is an imaging unit used to acquire images of the wound edge region and participate in geometric calibration, and its calibration process is also included. By calibrating the camera's intrinsic and extrinsic parameters, installation angle, and the relative position of the imaging window / airbag, a one-to-one correspondence between "pixel coordinates and inherent coordinates of the wound edge" is established.

[0019] Among them, the transparent airbag is a transparent elastic cavity covering the wound, which serves as a medium for applying pressure and isolating the outside world. Its transparency allows the camera to observe the wound edge, and the elastic deformation enables controllable pressure application. The geometric / optical properties (curvature, refraction) are compensated by the algorithm after calibration to avoid the influence of imaging distortion on the measurement.

[0020] Among them, the optical geometric relationship is the geometric and refractive mapping of the overall imaging link from camera to imaging window to transparent airbag to wound edge surface, including intrinsic / extrinsic parameters, distortion model and cross-medium (air / material) refractive model; this relationship is used to restore the original image to the inherent coordinates of the wound edge and eliminate systematic errors introduced by the vehicle and medium.

[0021] Among them, the safe zone of the wound edge is the area of ​​the wound edge (ROI) that is jointly determined by doctors and algorithms at the baseline stage and is suitable for measurement and stress. It excludes high-risk / low-confidence areas such as necrosis, exudation, reflection and edge tearing, and is used as a unified analysis range for subsequent displacement-strain calculation, threshold determination and visualization.

[0022] Among them, the baseline imaging data is a collection of wound edge images acquired and denoised and registered after optical geometric calibration before the rewarming chamber is closed or pressurized. It records the reference state of "zero pressure and zero heat" and contains ROI masks and coordinate mappings.

[0023] Among them, test pressurization is a small, controlled increase in pressure (generally slow, short-term, and with a basically constant temperature) applied in the early stage of treatment to detect tissue response. Its purpose is to induce a measurable deformation response within safe boundaries.

[0024] Among them, continuous imaging data are time-series images acquired at a fixed frame rate from the safety zone of the blast edge during test pressurization and subsequent procedures.

[0025] Among them, the surface displacement strain time sequence field is the displacement vector field and the corresponding surface strain field sequence that are calculated in the inherent coordinate system of the edge and change with time. The result has corrected for the medium distortion and non-pressurized components and reflects the real tissue response "caused by pressurization".

[0026] Among them, the stepped pressurization is a graded pressurization protocol implemented to obtain steady state and compliance (equal steps, pressure holding at each level and stability judgment); the steady state statistical points of each step are used to fit the pressure-strain relationship and identify the transition interval from "easy to difficult to compress", avoiding viscoelastic deviation caused by continuous climbing.

[0027] Among them, wound tissue compliance data is a set of tissue pressure response information obtained by step-up pressure sampling and fitting, including steady-state characteristics at each level, structured description of the fitted curve and (optional) evidence of the transition interval; it is used to formulate the initial target pressure zone and subsequent pressure-temperature coupling strategy.

[0028] The pressure-temperature coupling cycle is a closed-loop execution process that uses compliance data as a reference and, under the premise of ensuring a safe deformation envelope, links and adjusts the pressure and zone temperature, while monitoring deformation and quality indicators in real time. Its output includes temperature uniformity, strain uniformity, and risk event records during the maintenance period.

[0029] Among them, wound tissue maintenance data is a running data package collected and summarized during the stable phase of the pressure-temperature coupling cycle, reflecting the long-term maintenance effect and boundary margin within the target band; it provides a basis for perturbation correction and long-term adaptation (e.g., whether hotspot suppression is effective, whether there is a risk of penetration).

[0030] Among them, pressure-temperature coupled perturbation correction is a process of superimposing low-amplitude, low-frequency pressure / power perturbations near the target and correcting the target setting and partition weights by observing changes in tissue sensitivity; it is used to verify whether the steady state is truly stable and to make small-step corrections to potential drift (edema, changes in body position).

[0031] Among them, wound tissue stability data is the final treatment setting and stability evidence set formed after perturbation correction convergence, including remaining safety margin, sensitivity spectrum, partition weighting / masking suggestions and abnormal rollback rules.

[0032] Among them, the deformation visualization data is a visualization output (strain heatmap, displacement vector superposition, wound edge circumferential 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 accompanied by timestamps and set parameter watermarks for clinical interpretation, efficacy tracking and auditing.

[0033] Specifically, with the wound rewarming chamber open, a calibration target (or the natural texture of the wound edge) is placed to calibrate the optical geometry of the pose calibration camera and the transparent airbag. This calibration simultaneously calibrates the camera's intrinsic and extrinsic parameters, lens distortion, and the refractive model of the airbag / window, establishing a mapping from image to the inherent coordinates of the wound edge. Under fixed ambient light and exposure conditions, ≥N frames of baseline images are acquired and temporally denoised. The "safe area of ​​the wound edge" (ROI) is automatically / semi-automatically delineated and fixed. A baseline imaging data package is generated, containing the inherent coordinate grid, refraction / distortion correction parameters, and the ROI mask.

[0034] Server 104 controls the wound rewarming chamber to initiate a small-scale "probing pressurization" (slow rise, constant temperature) to test pressurize the wound. During the test pressurization, continuous images of the safe area of ​​the wound edge are acquired at a fixed frame rate. The correlation matching (DIC) and optical flow displacements are calculated in parallel within the intrinsic coordinate system. Then, a "pressureless counterfactual" trajectory is generated using the baseline frame to remove the non-pressurized components and fuse them to obtain the sub-pixel displacement time series. The displacement field is topologically and physically consistent with strain reconstruction (boundary conservation, radial compliance monotonicity) to form a surface displacement strain time series field, and quality indicators (flattening trend, low fluctuation, no hot spot expansion) are output simultaneously.

[0035] Server 104 controls the wound rewarming chamber to execute a stepped pressurization protocol (equal steps + pressure holding window) for stepped pressurization. During the step pressurization process, at the end of each pressure holding stage, stability is automatically determined according to the triple criteria of "very small slope, very low variance, and no hot spot expansion", and representative strain / displacement statistics of that stage are extracted. Individualized pressure-strain compliance curves are fitted across each stage to identify the transition interval from "easy to difficult to compress", forming wound tissue compliance data, which includes step-level steady-state evidence, fitting residuals, and candidate safe / effective pressure zones.

[0036] Using wound tissue compliance data (or existing pressure-temperature optimization data) as a reference, the wound rewarming chamber is controlled to perform pressure-temperature coupling cycles on the wound. During the execution of the pressure-temperature coupling cycle, the pressure is adjusted in small steps within the target zone, and the temperature is uniformly PWM-based in each zone. The controller compares the safe deformation envelope (SDE) with the quality indicators in real time, and automatically reduces pressure / power when approaching the boundary. The strain uniformity, hot spot suppression effect, temperature consistency, and event logs are continuously recorded during the "maintenance period" and summarized into wound tissue maintenance data.

[0037] Low-amplitude, low-frequency pressure perturbations (or slight partitioned power perturbations) are superimposed near the wound tissue maintenance data. The sensitivity and stability of the strain time series to the perturbations are observed. If the sensitivity of hot spots increases or the connectivity tends to be connected, the target pressure is reduced / the corresponding partition is shielded and the temperature is re-homogenized. The process is iterated until the sensitivity stabilizes, the hot spots do not expand, and the boundary margin is sufficient. The wound tissue stability data is then obtained, which includes the final settings, partition weight reduction / shielding suggestions, remaining safety margin, and sensitivity spectrum.

[0038] The surface displacement-strain time series, steady-state window, and stability data in the wound tissue stability data are synchronized and registered with the inherent coordinates of the wound edge according to a unified time axis. After determining the final target pressure / temperature and the corresponding steady-state range, the displacement vector and the circumferential curve of the wound edge are superimposed on the strain thermogram as the main layer, and the safety deformation envelope, penetration risk and hot spot mask and key event markers are superimposed. At the same time, steady-state evidence, dual-flow consistency residuals and uncertainty prompts are solidified as evidence layers for each frame. After color level standardization and scale annotation, a replayable dynamic image and an archiveable static snapshot (including timestamp and set parameter watermark) are output to form rewarming chamber deformation visualization data that can be used for clinical interpretation, efficacy tracking and auditing.

[0039] In the aforementioned method for visually monitoring deformation during the pressurization process of a wound rewarming chamber, the optical geometric relationship between the camera and the transparent airbag is first established using pose calibration while the wound rewarming chamber is not closed. This allows for the acquisition of baseline imaging data of the safe area at the wound edge, ensuring non-invasiveness and safety during the imaging process. Subsequently, the wound is subjected to test pressurization under control, and the surface displacement strain time series field is calculated based on the continuous imaging results, enabling the dynamic and precise capture of minute deformations of the wound tissue. On this basis, a step-pressure test is further implemented to obtain tissue compliance data, providing a scientific basis for evaluating the biomechanical response of the wound under different external forces. Then, a pressure-temperature coupled cyclic test is combined to obtain tissue maintenance data, reflecting the functional recovery capacity of the wound tissue under comprehensive stress conditions from multiple dimensions. Finally, through pressure-temperature coupled perturbation correction, data visualization results that can intuitively demonstrate wound stability are generated. It not only enables the systematic and quantitative assessment of key parameters such as wound tissue strain characteristics, compliance, and stability, but also provides a scientific and reliable basis for the development of personalized treatment strategies in clinical practice, greatly improving the safety, specificity, and controllability of wound treatment, and effectively increasing the success rate of treatment for injured patients in cold regions.

[0040] In an exemplary embodiment, based on wound tissue compliance data, a wound rewarming chamber is controlled to perform pressure-temperature coupled cycling on the wound to obtain wound tissue maintenance data, including steps 302 to 308. Wherein:

[0041] Step 302: Perform nonlinear fitting on the wound tissue compliance data to obtain wound tissue fitting data.

[0042] Step 304: Based on the wound tissue fitting data, perform a dual-threshold joint analysis of the wound treatment to obtain wound pressure and temperature optimization data.

[0043] Step 306: Using the wound pressure and temperature optimization data as a reference, perform dynamic analysis of the wound pressure and temperature cycle to obtain the wound pressure and temperature limit data.

[0044] Step 308: Based on the wound pressure and temperature optimization data and the wound pressure and temperature limit data, control the wound rewarming chamber to perform pressure and temperature coupling circulation on the wound to obtain wound tissue maintenance data.

[0045] Nonlinear fitting involves first performing steady-state screening and hysteresis-reducing shaping on the compliant samples obtained from "step-up compression," and then performing curve fitting under the physical shape priors such as monotonicity, amplitude limiting, and fast-then-slow, as well as the sparse constraints of "minimizing inflection points," to obtain an individualized curve that truly reflects the "pressure-deformation" relationship, rather than a simple least-squares empirical curve.

[0046] Among them, the wound tissue fitting data is a set of structured results generated by the above nonlinear fitting process and can be directly called by subsequent steps. It includes at least: fitting results for representing individualized "pressure-deformation" relationships, corresponding error / uncertainty information, and evidence indexes related to sample steady state.

[0047] Among them, the treatment dual-threshold joint analysis is a joint judgment process based solely on "wound tissue fitting data": on the one hand, it identifies the effective hemostasis threshold (EHST, interval definition) as the compliance curve transitions from "easy to compressible" to "difficult to compress", and on the other hand, it generates a spatiotemporally joint safe deformation envelope (SDE) based on the deformation space statistics and uncertainty boundary of the fitting data; the two thresholds serve as hard constraints to jointly limit the treatment feasibility domain, avoiding empirical bias of single-point thresholds.

[0048] Among them, the wound pressure-temperature optimization data is an execution setting set obtained through robust optimization of multiple objectives (hemostasis effect, deformation uniformity, energy / temperature rise constraint) under the dual constraints of EHST and SDE. It usually includes target pressure, target temperature, zone heating weight and maintenance period perturbation strategy. This data transforms "what can be done" into "how should be done", which is used to drive the subsequent closed loop and simulation.

[0049] Among them, the dynamic analysis of pressure-temperature cycle uses wound pressure-temperature optimization data as a reference and combines the uncertainty in the fitted data to perform adversarial (most unfavorable case) time domain scanning and boundary tracking on cycle templates such as pressure increase, constant pressure and temperature increase, maintenance period perturbation and pressure decrease. It identifies when events such as touching the SDE boundary, hot spot merging or penetration approach, and quantifies the allowable rate, residence time and safety margin.

[0050] Among them, the wound pressure and temperature limit data is an executable constraint list and guard rules output by the pressure and temperature cycle dynamic analysis. It at least clarifies the maximum / minimum acceptable rate of rise and fall, minimum residence time, perturbation amplitude / frequency range, zonal weight reduction or shielding suggestions, and prohibited pressure and temperature combinations for each stage. Its role is to act as a "hard guardrail" in actual closed-loop control to ensure that the optimization objectives are implemented within the safe and feasible domain.

[0051] Specifically, firstly, based on the "pressure-holding steady-state evidence" of each step in the wound tissue compliance data, transitional samples are removed and hysteresis segments are hysteresis-reduced to make the samples equivalent to path-independent steady-state points. Then, the data processed in the previous step is subjected to nonlinear fitting under constraints such as monotonicity, amplitude limiting, fast-then-slow shape priors, and minimum inflection segments (sparse inflection points), outputting an individualized pressure-deformation compliance curve. At the same time, error bands, steady-state window indexes, and structured fields for subsequent judgments are generated to match the curve, thus forming wound tissue fitting data.

[0052] Robust identification of transition intervals in the wound tissue fitting data is performed, defining the interval "from easily compressible to difficult to compressible" as the effective hemostasis threshold (EHST). Furthermore, a spatiotemporally joint safe deformation envelope (SDE) is generated using deformation spatial statistics and uncertainty boundaries within the wound tissue fitting data to exclude combinations that may result in patchy or penetrating hemostasis. Finally, using EHST and SDE as hard constraints, a multi-objective robust optimization (efficacy improvement, deformation uniformity, energy constraint) is constructed to obtain target pressure, target temperature, zonal heating weights, and perturbation strategies, forming wound pressure-temperature optimization data.

[0053] Wound pressure-temperature optimization data is mapped to a "pressure-temperature feasible channel" on the time axis. Based on the uncertainty range given by the wound pressure-temperature optimization data, adversarial counterfactual scanning is performed on cyclic templates such as pressure increase, constant pressure and temperature increase, maintenance period perturbation and pressure decrease to track the earliest time and conditions of events such as triggering SDE boundary, hot spot clustering or penetration convergence. On this basis, the feasible channel is safely tightened and rhythm is defined to generate the maximum / minimum allowable rise / fall rate, minimum residence time, perturbation amplitude / frequency range and prohibited pressure-temperature combination list for each stage, forming wound pressure-temperature constraint data.

[0054] Server 104 sets targets according to wound pressure and temperature limits and executes closed-loop operations under the constraints of these limits. This includes adjusting pressure in small steps within the target zone, uniformizing temperature by zone, monitoring deformation timing and safety envelope distance online, and immediately reducing pressure, power, or shielding zones according to defined rules and recording rollbacks when a trigger event occurs. During the maintenance period, low-amplitude perturbations are periodically applied to verify perfusion / deformation stability and fine-tune the adjustment law parameters. Ultimately, operational indicators such as target achievement, hotspot suppression effect, boundary margin, perturbation response, and abnormal handling logs are collected and summarized into wound tissue maintenance data.

[0055] In this embodiment, nonlinear fitting is used to first eliminate transition and hysteresis biases to obtain a stable individual response. Furthermore, dual thresholds (EHST for efficacy + SDE for safety) are used as hard constraints to generate pressure-temperature optimization data, ensuring both hemostasis and tissue safety from the outset. Dynamic cycle analysis provides pre-defined combinations of acceleration / deceleration rates, minimum residence time, and disablement, reducing the risks of edge contact, penetration, and overheating. Finally, a contractual closed-loop execution based on "optimization + constraint" is implemented, with perturbations used to verify steady-state stability, forming traceable and maintainable data. This approach enables faster threshold attainment, more stable maintenance, lower concurrency, and higher repeatability / auditability.

[0056] In an exemplary embodiment, based on wound tissue fitting data, a dual-threshold joint analysis of the wound treatment is performed to obtain wound pressure-temperature optimization data, including steps 402 to 410. Wherein:

[0057] Step 402: Perform structured analysis on the wound tissue fitting data to obtain compliance curve data, deformation field statistics, fitting parameter error range, and pressure holding steady-state evidence.

[0058] Step 404: Window filtering is performed on the pressure-holding steady-state evidence to obtain a subset of steady-state data.

[0059] Step 406: Based on the compliance curve data and the error range of the fitting parameters, analyze the turning interval of the wound to obtain the effective hemostatic threshold pressure data.

[0060] Step 408: Based on the deformation field statistics, the error range of the fitting parameters, and the steady-state data subset, perform spatiotemporal connectivity stability analysis on the wound to obtain safe deformation envelope data.

[0061] Step 410: Perform robust optimization on the effective hemostasis threshold pressure data and the safe deformation envelope data to obtain wound pressure-temperature optimization data.

[0062] Structured analysis is a process of organizing and decomposing "wound tissue fitting data" under the inherent coordinates of the wound edge and a unified time axis. It transforms the original / intermediate results into a standard view that can be directly called by subsequent algorithms. It usually produces four types of core elements simultaneously: compliance curve data, deformation field statistics, fitting parameter error range and pressure holding steady state evidence. It also establishes an index relationship between these elements and metadata such as step pressure, timestamps, and ROI to avoid redundant calculations and ensure traceability and consistency.

[0063] Among them, compliance curve data is an individualized function / piecewise representation of the pressure-deformation (or strain) relationship and its sampling point mapping. It is derived from the steady-state statistical points of each step pressure level and fitted with shape prior constraints. It is used to identify the transition behavior from "easy to compressible" to "difficult to compress" and is the direct basis for determining the effective hemostasis threshold and formulating the target pressure zone.

[0064] Among them, the deformation field statistics are spatial-temporal statistics obtained by summarizing the surface displacement-strain time series field within the steady-state window of each step. Typical statistics include circumferential average strain, maximum principal strain, number / area / maximum span / duration of hot spot connected domains and minimum distance from the creation edge, etc., which are used to assess whether they are in patches, whether they are connected and their evolution trend.

[0065] Among them, the fitting parameter error range is the uncertainty bound obtained by compliance fitting and related mapping after residual analysis, cross-validation or scene sampling. It is used to set robust thresholds and safety margins for threshold identification, connectivity analysis and optimization solution, so as to avoid misjudgment and out-of-bounds caused by point estimation.

[0066] Among them, the steady-state evidence for pressure holding is the joint criterion and timestamp record of the strain curve “flattening trend”, the fluctuation “continuously converging” and the hot spot “not expanding spatially” within the short window of the pressure holding stage. When necessary, it is supplemented by pressure stability and visual quality guards to confirm which frames / intervals can be used as reliable samples to enter the fitting and safety analysis.

[0067] Among them, the steady-state data subset is the time period and its corresponding pressure level and statistical set retained after screening the steady-state evidence of pressure holding. It serves as the sole time window source for subsequent transition interval identification (EHST) and spatiotemporal connectivity stability analysis (SDE), ensuring that all judgments are based on highly reliable data.

[0068] The transition interval is the pressure section on the compliance curve that transitions stably from "easy to compressible" to "difficult to compress." It is usually detected by a combination of a significant decrease in slope and a curvature transition, along with uncertainty verification and steady-state backtracking confirmation.

[0069] Among them, the effective hemostasis threshold pressure data is the pressure range result output after robustness and repeatability verification based on the transition interval, with boundary confidence and source step index attached, which is used to constrain subsequent optimization and control so that the treatment is "sufficient to stop bleeding" while avoiding excessive compression.

[0070] Among them, the spatiotemporal connectivity stability analysis is a joint process that applies adaptive thresholding, high strain mask generation, directional penetration / percolation testing, morphological counterfactual robustness testing and durability analysis to the strain field within the steady-state window, in order to quantify whether hotspots are formed, whether they tend to penetrate, and their sensitivity to uncertainty.

[0071] Among them, the safety deformation envelope data is a set of results output according to the pressure level, including a space safety mask, a risk mask, and a conservative buffer zone. It can also include partition weight reduction / masking suggestions as a "hard constraint" for optimization and control, ensuring that the pressure-temperature strategy always operates within the safe and feasible domain.

[0072] Among them, the robust optimization process is a multi-objective solution process carried out under the dual constraints of effective hemostasis threshold and safe deformation envelope, combined with uncertainty delimitation, with the goal of improving efficacy and deformation uniformity, and suppressing risk and energy load.

[0073] Specifically, the wound tissue fitting data is structured and analyzed under the inherent coordinates of the wound edge and a unified time axis to obtain compliance curve data (the function / piecewise representation of pressure-deformation and corresponding sampling points), deformation field statistics (circumferential average strain, maximum principal strain, area / span / duration of hot spot connected regions, etc. summarized by steps / time), fitting parameter error range (uncertainty interval / envelope formed based on fitting residuals and cross-validation), and pressure-holding steady-state evidence (the judgment results of trend flattening, fluctuation convergence, and hot spot non-expansion, as well as their timestamps, pressure levels, and other metadata).

[0074] According to the "three steady-state criteria" (strain increment tends to flatten, short-window fluctuations converge, and hotspots do not expand), the pressure holding intervals of each order in the pressure holding steady-state evidence are automatically judged for stability and denoised. Frame segments containing abnormal residuals or occlusions are removed, and those that pass are labeled with steady-state tags and confidence levels. The resulting data will be aggregated through the judged stable frames along with their corresponding pressure levels, deformation statistics, and uncertainty labels to form a steady-state data subset.

[0075] Under a unified time axis and inherent coordinates of the wound edge, the sampling points corresponding to the subset of steady-state data are mapped onto the compliance curve data. Multi-scale segmented fitting and smoothing are performed based on the shape prior of minimum inflection points and monotonic amplitude limiting. Then, candidate transition zones from "easy to difficult to compress" are screened out using the slope decay-curvature transition joint criterion. For each candidate zone, uncertainty verification is performed in combination with the error range of the fitting parameters (one-order cross-validation, neighborhood consistency, and abnormal step removal). The persistence and repeatability of its start and end positions are verified by the steady-state evidence backtracking kernel, thereby compressing it into an intervalized effective hemostasis threshold. Finally, the effective hemostasis threshold pressure data are output, which gives the transition interval, its confidence label, and the source step index, as direct input for subsequent safety envelope analysis and robust optimization.

[0076] Under the inherent coordinate system of the origin, the strain field within the steady-state data subset is adaptively segmented and thresholded for uncertainty based on the deformation field statistics and the error range of the fitting parameters, generating a high-strain mask sequence. Simultaneously, under the constraint of the origin-peripheral boundary, the sequence is subjected to directional penetration / seepage determination and the minimum fracture scale is evaluated. The temporal trajectory of the number, area, maximum span, and residence time of connected domains is recorded simultaneously. Furthermore, within the uncertainty range, the mask is subjected to counterfactual robustness tests for morphological expansion / corrosion and random pose perturbation. Occasional short-term hotspots are eliminated by persistence analysis, resulting in a penetration risk map, a robustness score map, and a residence expansion evidence map. Finally, the above layers are fused and cascaded according to pressure levels, outputting graded indexed safety deformation envelope data, clarifying the spatial safety mask, risk mask, conservative buffer zone, and recommended weighted / shielded partitions.

[0077] Using the effective hemostatic threshold pressure range (EHST) and safe deformation envelope data (SDE) as hard constraints, a multi-objective optimization model with sub-Bruker bars is constructed in conjunction with the fitting parameter error range. Under the synergy of scenario sampling and constraint relaxation penalty, the model simultaneously improves therapeutic efficacy and deformation uniformity / temperature balance, suppresses hot spot risk and energy load, and implements feasible domain regression for infeasible solutions. The decision variables include target pressure, target temperature, zone heating weights, and maintenance period perturbation amplitude and frequency. After solving, the model outputs wound pressure and temperature optimization data that can be directly deployed, which specifies each set value, zone weight reduction / masking suggestions, minimum boundary margin for SDE, and scenario coverage, as well as other executable and auditable indicators.

[0078] In this embodiment, structured analysis and window filtering are used to retain only steady-state and traceable data sources, avoiding fitting bias caused by viscoelastic transitions. Furthermore, an effective hemostasis threshold (EHST) is obtained by replacing single-point thresholds with "turning point intervals," combined with the generation of a safety deformation envelope (SDE) through spatiotemporal connectivity stability analysis, balancing efficacy and tissue safety. Robust optimization is performed within the uncertainty boundary, directly producing individualized pressure and temperature settings with safety margins. This enables faster threshold attainment, lower concurrency (avoiding the risk of extensive / penetrating hemostasis), higher robustness and repeatability, and significantly reduces manual intervention and trial-and-error costs.

[0079] In an exemplary embodiment, spatiotemporal connectivity stability analysis of the wound is performed based on deformation field statistics, fitting parameter error range, and steady-state data subset to obtain safe deformation envelope data, including steps 502 to 512. Wherein:

[0080] Step 502: Perform a joint analysis of the trend flattening, fluctuation convergence, and spatial non-expansion of the candidate steady-state periods labeled in the steady-state data subset to determine the steady-state window of the pressure-holding segment.

[0081] Step 504: Based on the deformation field statistics and the error range of the fitting parameters, perform adaptive segmentation analysis 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.

[0082] Step 506: Based on the strain mask sequence and uncertainty envelope, perform directional penetration / seepage analysis on the wound to obtain penetration risk data and safety margin data.

[0083] Step 508: Perform counterfactual morphological stress tests on the high-strain mask sequence and the uncertainty envelope to obtain robustness analysis data; Step 510: Based on the steady-state data subset and strain mask sequence, perform time series analysis on the area, maximum span, and duration of occurrence of the strain connectivity domain to obtain the hotspot residence expansion evidence map. Step 512: Based on the risk data, safety margin data, robustness analysis data, and hotspot residence and expansion evidence map, the wound is subjected to cascaded threshold processing to obtain safety deformation envelope data.

[0084] Among them, the trend flattening is when the rate of change of the circumferential average strain / maximum principal strain with time decreases significantly and remains in a very small range within a short time window, which is manifested as the slope of the curve being close to zero. This is used to determine whether the deformation process has entered the stable stage from the transition stage.

[0085] Among them, fluctuation convergence is characterized by a continuous decrease in the variance / variance coefficient of the strain sequence within the same short window without any sharp peaks, indicating that the signal fluctuates slightly around a stable mean.

[0086] Among them, the lack of spatial expansion means that the number, area and maximum span of the connected domains in the high-strain region within the short window do not continue to increase, and no new clusters or connections appear, indicating that the risk hotspots have not expanded.

[0087] Among them, the steady-state window of the pressure holding section is a continuous time segment that simultaneously satisfies "trend flattening, fluctuation convergence, and no spatial expansion" under the pressure holding condition (and is protected by pressure / visual guard).

[0088] Among them, adaptive segmentation analysis adaptively sets the threshold and scale based on the fitting error and data distribution, and processes the strain field separately in different intensity ranges to improve the robustness of segmentation and statistics.

[0089] Among them, the strain mask sequence is a high strain region sequence obtained by adaptively thresholding the strain field of each frame within a steady-state window.

[0090] Among them, the strain connectivity domain is a connected high-strain component marked according to a preset neighborhood rule in a single-frame strain mask, which can be tracked in time to statistically measure attributes such as area, span and duration of occurrence.

[0091] Among them, the uncertainty envelope is a pixel-level probability band (stabilizing kernel / boundary band / exclusion band) formed by frequency statistics of multiple instance masks under perturbations such as thresholding, calibration and reconstruction, which is used to quantify the confidence range of the segmentation results.

[0092] Among them, the directional penetration / permeation analysis examines whether the high-strain mask forms a penetration path from the wound edge to the periphery under the constraint of the wound edge-peripheral boundary, and evaluates its forming conditions and path characteristics.

[0093] Among them, the breakthrough risk data is the risk quantification result output by the directional seepage analysis, including whether it is breakthrough, the earliest trigger time, path length / width, etc., which is used for the first-level safety determination.

[0094] Among them, the safety margin data are measures such as the minimum fracture scale and minimum shielding width required to break the potential breakthrough in the case of non-breakthrough, reflecting the remaining safety space from the risk boundary.

[0095] Among them, the counterfactual morphological stress test involves applying multi-scale expansion / erosion and pose perturbation to the mask within an uncertainty range, observing changes in connectivity and clumping tendency, in order to assess the robustness of the morphology under perturbation.

[0096] Among them, the robustness analysis data are quantitative scores and labels obtained from counterfactual morphological stress tests, which are used to characterize the tendency of hotspots to maintain or form risk patterns under disturbances.

[0097] Among them, the hotspot residence expansion evidence map is an evidence layer formed by accumulating the occurrence duration (residence) and area / span growth trend (expansion) of connected domains in space after time-series tracking, and is used to identify "contiguous and continuous" risk areas.

[0098] Among them, cascaded threshold processing involves thresholding and merging multiple layers such as penetration risk, safety margin, robustness score and residency expansion evidence according to priority, eliminating unsafe areas and smoothing boundaries, thereby generating safe deformation envelope data.

[0099] Specifically, the candidate steady-state time periods labeled in the steady-state data subset are aligned with the pressure time series. Within the inherent coordinates of the creation edge, short-window slope / rate of change assessment (trend flattening) and short-window variance / coefficient of variation assessment (fluctuation convergence) are performed on the circumferential average strain and the maximum principal strain. Short-window comparison is also performed on the number, area, and maximum span of connected domains in the high-strain region (no spatial expansion). Each threshold is adaptively set according to the error range of the fitting parameters, and pressure stability and visual quality protection (no abrupt changes, no occlusion / reflection) are superimposed. The segment that meets the three criteria and is continuously maintained for ≥ preset duration is identified as the steady-state window of the pressure holding segment.

[0100] Within the steady-state window of each pressure-holding segment, an adaptive segmentation threshold is first applied to the strain field based on deformation field statistics to obtain an initial strain mask sequence. Subsequently, Monte Carlo / bootstrapping perturbations (including threshold expansion, displacement field noise reconstruction, calibration perturbations, etc.) are applied to the threshold and displacement-strain reconstruction link based on the fitting parameter error range to generate the target strain mask sequence. Pixel-level occurrence frequency statistics are performed on the target strain mask sequence to form a confidence map, and a stable kernel, exclusion zone, and boundary zone are obtained by double-threshold segmentation. The stable kernel is marked with connectivity to output the strain connectivity domain. The boundary zone is morphologically calibrated (small-scale expansion / erosion and bridging removal) along the creation-periphery direction and then encapsulated into an uncertainty envelope (containing pixel-level probability and bandwidth indicators).

[0101] Based on strain mask sequences and uncertainty envelopes, a directional seepage map is constructed under the constraint of the wound edge-peripheral boundary. For each pressure level / window, it is detected whether there is a high-strain penetration path from the wound edge to the periphery. If it exists, the penetration risk data (penetration mark, earliest trigger time, path length / width) is output. If it does not exist, the minimum fracture scale and minimum shielding width required to break the penetration are calculated as safety margin data.

[0102] Based on the strain mask sequence, multi-scale expansion / erosion and random pose perturbation (translation / rotation / scale) are performed on the mask within the allowable range of the uncertainty envelope. The changes in connectivity, clumping rate and path forming rate are statistically analyzed to form a robustness score for each connected region and the global mask. The higher the score, the easier it is to maintain or form a risky pattern under perturbation. The output is robustness analysis data and backfilled into the corresponding stress level / window.

[0103] Based on the steady-state data subset index, the strain connected regions of the strain mask sequence are time-series tracked (associated according to the overlap / proximity criterion), and the occurrence duration (dwelling), area and time derivative of the maximum span (expansion trend) of each connected region are calculated. Dwelling and expansion evidence maps aggregated by pixels / grid are generated. This evidence map is used to characterize the risk features of "patchy and continuous".

[0104] The risk data, safety margin data, robustness analysis data, and hotspot retention and expansion evidence map are fused into multiple layers at the same pressure level. The multi-layer fusion uses "whether it is connected" as the first-level hard threshold, followed by "safety margin ≥ threshold" and "robustness score ≤ threshold" as the second-level thresholds, and finally "retention / expansion evidence below threshold" as the third-level threshold. Unsafe / edge areas are removed from the inside out according to the rules. Morphological smoothing is performed on the connected areas and a conservative buffer band is superimposed. The output is safety deformation envelope data (including spatial safety mask, risk mask, and recommended downweight / shielding partitions) indexed by pressure level.

[0105] In this embodiment, the steady-state window of the pressure-holding segment is precisely locked by a joint analysis of "trend flattening + fluctuation convergence + spatial non-expansion," avoiding bias introduced by transitional data. Secondly, adaptive segmentation and the construction of an uncertainty envelope yield a more robust strain mask sequence and strain connectivity domain. Then, directional penetration / seepage analysis upgrades "whether it is penetrated" to a primary hard criterion and quantifies the safety margin, replacing the traditional single-point threshold. Furthermore, counterfactual morphological stress testing and evidence of hotspot residence and expansion identify "contiguous and persistent" high-risk areas. Finally, cascaded threshold fusion of multi-layer evidence outputs safe deformation envelope data that can be directly used for control. This significantly reduces the false positive rate and the risk of penetration / overpressure, improving the quantifiability of individualized safety boundaries, algorithm robustness, and reproducibility of clinical implementation.

[0106] In an exemplary embodiment, wound pressure-temperature optimization data is used as a reference condition to perform dynamic analysis of wound pressure-temperature cycles, obtaining wound pressure-temperature limiting data, including steps 602 to 606. Wherein:

[0107] Step 602: Based on the wound pressure-temperature optimization data and the safety deformation envelope data, perform time-domain channel analysis on the pressure-temperature parameters of the wound rewarming chamber to obtain feasible pressure-temperature channel data.

[0108] Step 604: Based on the feasible pressure-temperature channel data and the error range of the fitting parameters, perform adversarial counterfactual scanning analysis 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.

[0109] Step 606: Perform safety rhythm delimitation processing on the loop boundary trajectory and the set of triggering events to obtain wound pressure and temperature limit data.

[0110] Among them, the time-domain channel analysis maps the wound pressure-temperature optimization data (target pressure / temperature and partition weights, perturbation strategy) onto a cyclic template (pressure increase - constant pressure and temperature increase - maintenance - pressure decrease) on a unified time axis, and intersects it with the safety deformation envelope data hourly to obtain the pressure band, temperature band and boundary margin available in each partition at each time / stage. At the same time, it solidifies the state guard and transition conditions to form the analysis and processing of the feasible search space for subsequent scans.

[0111] Among them, the pressure-temperature feasible channel data is the result set of time-domain channel analysis, which records information such as "allowable pressure / temperature range, partition weighting / masking mark, minimum safety margin, status guard and transition conditions" by time × partition; it transforms "optimization settings + safety envelope" into time-series feasible channels that can be directly traversed by the algorithm.

[0112] Among them, the adversarial counterfactual scanning analysis constructs the most unfavorable perturbation (threshold drift, calibration perturbation, power deviation, measurement noise, etc.) within the search space of feasible pressure and temperature channel data, combined with the error range of fitting parameters. It then scans each sub-template of pressure boosting, constant pressure heating, maintenance period perturbation and pressure reduction step by step to find the path and time that first touches the safety boundary (SDE) or triggers the risk mode.

[0113] Among them, the cyclic boundary trajectory is the "nearest boundary point-time" sequence formed when the system state approaches the safety boundary along the feasible path during the adversarial counterfactual scanning process. It includes the corresponding pressure / temperature / zoning combination, minimum safety margin and normal crossing trend, and is used to quantify "how far away from the risk" into executable boundary evidence.

[0114] The trigger event set is a set of discrete events that are determined to have reached or approached the risk threshold during the scan, such as SDE edge contact, penetration approach, hotspot clustering, rate over-limit, excessive temperature rise, quality guard mismatch, etc. Each event carries a timestamp, location / partition, trigger condition, and associated boundary trajectory index.

[0115] Among them, the safety rhythm delimitation process is the process of translating "cyclic boundary trajectory + trigger event set" into executable timing constraints and guard rules. It outputs the maximum / minimum rise / fall rate, minimum dwell time, allowable perturbation amplitude / frequency, partition weight reduction / masking and prohibition combination list for each stage, and solidifies the automatic backoff / recovery strategy as a hard guardrail and scheduling contract for closed-loop control.

[0116] Specifically, the target pressure / temperature, zone heating weights, and perturbation strategies in the wound pressure-temperature optimization data are mapped to a unified time axis cyclic template (pressure increase - constant pressure heating - maintenance - pressure reduction), and intersected with the safety deformation envelope data hourly to form the feasible interval or polygon constraint set (allowable pressure band, allowable temperature band, zone weight reduction / masking mark and boundary margin) in each zone at each time / stage. At the same time, the state guards (touch threshold, visual / sensor quality guards) and transition conditions are recorded, and the pressure-temperature feasible channel data are summarized and output.

[0117] Using the feasible channel data of pressure and temperature as the search space, and combining the error range of the fitting parameters, the most unfavorable perturbation set (threshold drift, measurement noise, calibration perturbation, partition power deviation, etc.) is constructed. The four types of sub-templates, namely pressure boosting, constant pressure heating, maintenance period perturbation and pressure reduction, are scanned in the time domain segment by segment. In each time step, the nearest point and normal crossing trend of the feasible domain boundary are found. The timestamp / location of the triggering events such as the first touch of the SDE boundary, the breakthrough approach, the hot spot clustering, the rate exceeding the limit, and the excessive temperature rise, as well as the corresponding boundary trajectory (including the minimum safety margin and the triggering path) are recorded to form the cyclic boundary trajectory and the triggering event set.

[0118] The cyclic boundary trajectory and trigger event set are translated into executable constraints. These executable constraints include setting the maximum / minimum rise / fall rate and transition suppression slope for the rise / fall stage based on the earliest trigger boundary point; setting the minimum dwell time and allowable perturbation amplitude / frequency for the maintenance stage based on the event recurrence interval and margin decay; and generating a prohibition list and partition deweighting / masking suggestions for frequently triggered combinations. At the same time, the state guards (SDE edge, through signal, quality guard) and automatic rollback / recovery strategies are solidified, and the wound pressure and temperature limit data are finally output in stages × partition cataloging.

[0119] In this embodiment, feasible pressure-temperature channel data is obtained by intersecting the pressure-temperature optimization and the safety deformation envelope time-by-time, clarifying "what can be done" at each moment and partition. Then, adversarial counterfactual scanning is performed under fitting uncertainty to capture the cyclic boundary trajectory and trigger event set (edge ​​contact, penetration approach, clumping, overheating, rate over-limit, etc.) on the most unfavorable path in advance, transforming potential risks into quantitative evidence. Finally, hard barriers such as upper / lower limits of rise and fall rates, minimum dwell time, perturbation amplitude and frequency, and disabling combinations are generated through safety rhythm delimitation, and a backoff / recovery strategy is provided. This can significantly reduce the risk of over-limit and concurrency while ensuring efficacy, improve robustness, repeatability, and compliance, and reduce manual intervention and trial-and-error costs.

[0120] In an exemplary embodiment, nonlinear fitting is performed on wound tissue compliance data to obtain wound tissue fitting data, including steps 702 to 708. Wherein:

[0121] Step 702: Dehysteresis processing is performed on the wound tissue compliance data to obtain a steady-state equivalent sample set.

[0122] Step 704: Based on the steady-state equivalent sample set, perform nonlinear fitting processing of the pressure-deformation relationship with shape prior and topological sparsity constraints to obtain preliminary tissue fitting data.

[0123] Step 706: Perform fixation processing on the preliminary tissue fitting data to obtain wound fixation fitting data.

[0124] Step 708: Perform structured encapsulation processing on the wound solid boundary fitting data to obtain wound tissue fitting data.

[0125] Hysteresis removal is a preprocessing procedure that removes the effects of loading / unloading loops and viscoelastic tails from the pressure-deformation-time data collected by step compression. It involves aligning candidate steady-state periods, closing energy loops, correcting baseline drift, removing abrupt / occluded samples, and selecting representative points based on short window slope ≈ 0 and variance convergence, so that subsequent fitting is based on steady-state information without path dependence.

[0126] Among them, the steady-state equivalent sample set is the pressure and equivalent steady-state deformation pairs extracted by step pressure holding window after hysteresis treatment, along with their confidence and time / step index.

[0127] Among them, the pressure-deformation relationship is a function / piecewise mapping of the surface deformation (or strain) response of the tissue under external pressure under the inherent coordinates of the fossa and a unified time axis. It is used to characterize the mechanical compliance behavior of "easy to compress → difficult to compress" and is the basic object for threshold identification and control setting.

[0128] Among them, shape prior is a physical consistency constraint imposed on the pressure-deformation relationship, including monotonically non-decreasing, amplitude limiting (upper and lower bounds), fast at first and then slow (decreasing slope), and smooth continuity, etc., to avoid overfitting curves without physical meaning and to stabilize the identification of turning points.

[0129] Among them, topological sparsity constraint is a regularization strategy that minimizes the number and complexity of segments / inflection points when modeling curves (minimum number of inflection segments / minimum topological complexity). It suppresses spurious change points caused by noise while satisfying shape priors, making the model both concise and interpretable.

[0130] The initial fitting data is based on a steady-state equivalent sample set and is the initial nonlinear fitting data obtained under shape prior and topological sparsity constraints. It includes a continuously differentiable curve representation, its error / residual information, and candidate inflection point hints.

[0131] Among them, the fixed boundary processing is a verification process of performing bounded projection and most unfavorable disturbance back-projection on the preliminary fitting results under the constraints of equipment capacity and safety boundaries (such as safety deformation envelope, maximum allowable slope / curvature, etc.), prohibiting extrapolation out of bounds, forcing monotonic amplitude limiting, and calculating the remaining margin with respect to the safety boundary.

[0132] Among them, the wound solid boundary fitting data is a set of fitting results that meet both physical and safety constraints after solid boundary processing. It includes compatibility labels and minimum boundary margins and can be directly used for threshold identification, optimization and control without the need for safety verification again.

[0133] The structured encapsulation process involves organizing the wound solid-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 to obtain deformation / slope, inverse solution pressure, inflection zone query, and margin query) so that subsequent dual-threshold analysis, channel construction, and closed-loop control modules can directly consume them.

[0134] Specifically, in the original pressure-deformation-time series of wound tissue compliance data characterizing step pressure, the candidate steady-state periods are first aligned with the pressure guards to identify loading / unloading loops and viscoelastic tail regions; energy loop closure and baseline drift correction are performed on the transition segments to remove anomalous points such as jumps / occlusions, and representative points for each pressure holding step are selected 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 path-dependent steady-state equivalent sample set.

[0135] Under the conditions of monotonically non-decreasing, amplitude-limited (upper / lower bounds) and "fast-then-slow" shape prior, a piecewise smooth fitting (such as restricted splines / piecewise exponents or Gompertz / generalized logistic family) is performed on the steady-state equivalent sample set. The complexity is controlled by the "minimum number of turning points" sparsity regularization. The number of segments and smoothing intensity are automatically selected through cross-validation. The continuously differentiable fitting curve, segment boundary positions, residual spectrum and candidate turning point hints are output to form the preliminary fitting data.

[0136] Under the constraints of equipment capability and organizational safety boundaries, the preliminary fitting data of the tissue is subjected to bounded projection, which limits the maximum slope and curvature (Lipschitz and second-order upper bound), prohibits extrapolation beyond the boundaries, forces monotonicity and non-negativity, and re-projects the curves to the feasible region under the most unfavorable perturbation scenarios (threshold drift, calibration perturbation); the minimum boundary margin of the safety deformation envelope / equipment upper limit is calculated simultaneously and labeled with compatibility tags to obtain wound solidified fitting data that satisfies both physical and safety constraints.

[0137] The fixed wound fit data and its metadata are packaged into a callable interface, providing functions such as "deformation / slope / curvature given pressure, inverse pressure given deformation, inflection interval query, and boundary margin query". At the same time, audit information such as source step index, steady-state evidence reference, residual summary, version and timestamp are encapsulated and serialized into a standard format to form wound tissue fit data that can be directly consumed by dual-threshold joint analysis, time domain channel and subsequent control.

[0138] In this embodiment, hysteresis removal processing first generates steady-state equivalent samples without transition tails, significantly reducing fitting bias; nonlinear fitting under shape prior and topological sparsity constraints yields concise, interpretable, and anomaly-insensitive individualized pressure-deformation relationships; solid boundary processing projects the most unfavorable disturbance, equipment limits, and safe deformation boundaries onto the model, outputting an "executable curve" with built-in boundary margins, suppressing out-of-bounds and overpressure risks from the source; finally, it is structured and encapsulated into a standard interface, facilitating direct calls for subsequent dual-threshold analysis, time-domain channels, and closed-loop control, achieving higher accuracy and robustness, lower parameter costs, and stronger repeatability / traceability.

[0139] In an exemplary embodiment, the surface displacement strain time-series field of the wound is calculated based on continuous imaging data of the safe area around the wound edge, including steps 802 to 808. Wherein:

[0140] Step 802: Perform medium geometry reduction processing on the continuous imaging data of the wound edge safety area and the baseline imaging data to obtain the wound edge intrinsic coordinate frame sequence.

[0141] Step 804: Perform dual-stream counterfactual reduction on the original coordinate frame sequence to obtain the subpixel displacement temporal field.

[0142] Step 806: Perform topological physical reduction on the subpixel displacement temporal field to obtain the initial surface strain temporal field.

[0143] Step 808: Perform quality control reduction on the initial surface strain time series field to obtain the surface displacement strain time series field.

[0144] Among them, the medium geometry reduction process is to perform distortion correction, refraction correction and parallax compensation on the baseline and continuous imaging frame by frame under the constraints of pose calibration results and transparent airbag optical model (camera intrinsic and extrinsic parameters, lens distortion, cross-medium refraction / curvature). It also constructs inherent coordinates (such as polar coordinates / flattened grid) and unified ROI with the edge boundary as reference, so that all subsequent images fall on the same geometrically and optically consistent analysis plane.

[0145] Among them, the primordial coordinate frame sequence is a time series image set formed by uniformly mapping the baseline / continuous frames to the primordial coordinate system with the primordial edge as the origin or reference after medium geometric reduction processing. It already includes boundary mask, pixel-physical scale mapping and time alignment information.

[0146] Among them, the dual-stream counterfactual reduction process calculates the correlation matching (DIC) and optical flow displacement candidates in parallel on the intrinsic coordinate frame sequence of the creation edge. Combined with the "pressure-free counterfactual" trajectory generated by the baseline frame, the motion components unrelated to pressure are stripped off. The dual streams are dynamically weighted and sub-pixel refined by uniform residual gating, so as to retain only the real displacement components "induced by pressure".

[0147] The subpixel displacement temporal field is a displacement vector field sequence that is continuous on the time axis and spatially characterized with subpixel precision, obtained by dual-stream counterfactual reduction, and includes residual / confidence information for each frame (or each pixel).

[0148] Among them, the topological physical reduction process introduces physical-topological consistency constraints such as boundary conservation, radial compliance monotonicity, shear smoothing and topological preservation (no cracking / no pseudo-clustering) when converting the sub-pixel displacement field into a strain field. It also suppresses noise and non-physical gradients through regularized projection to ensure that the reconstructed strain matches the anatomical boundary and the actual force.

[0149] Among them, the initial surface strain temporal field is a spatiotemporally continuous surface strain field sequence that is consistent with the anatomical boundary of the wound, obtained after topological physical reduction processing.

[0150] Among them, the quality control specification processing is based on residuals and confidence thresholds, baseline alignment and visual quality guarding (occlusion / reflection / fogging / film texture interference), and performs low confidence frame removal, physically consistent temporal interpolation, color level / dimensional standardization and timestamp registration on the initial strain time series. The final output is a high confidence "surface displacement-strain time series field" that can be used for judgment and control.

[0151] Specifically, based on the pose calibration results and the optical geometric model of the transparent airbag (camera intrinsic and extrinsic parameters, lens distortion, cross-medium refraction / curvature), the continuous imaging data of the wound edge safety area and the baseline imaging data are first subjected to frame-by-frame distortion correction, refraction correction and parallax compensation. Then, the inherent coordinates (such as polar coordinates / flattened grid) are constructed with the wound edge boundary as a reference. The two types of images are uniformly mapped to the same wound edge inherent coordinate system and time synchronization and ROI mask solidification are completed. The resulting frame sequence of wound edge inherent coordinates is output after geometric-optical consistency.

[0152] On the inherent coordinate frame sequence, digital image correlation (DIC) and optical flow displacement candidates are calculated in parallel for adjacent frames (and with the baseline frame), and a dual-flow consistency residual map is generated. The pressureless counterfactual trajectory is inferred from the baseline imaging data, and background / breathing components that are not related to pressure are stripped away, leaving only pressure-induced motion. Then, the DIC / optical flow is dynamically weighted and fused and sub-pixel level sub-window is refined by residual gating to suppress single path deviation and noise drift, forming a physically consistent and temporally continuous sub-pixel displacement temporal field.

[0153] The subpixel displacement temporal field is converted into strain through a topologically and physically consistent strain reconstruction link. Under the priors of boundary conservation, radial compliance monotonicity, and shear smoothing, finite strain / small strain solution is used, combined with topological preservation (connectivity / non-fragmentation) and regularized projection (e.g., TV / Laplace constraints) to suppress crack-like artifacts and clumping distortion. Then, anomalous gradients are robustly suppressed and locally back-projected to output an initial surface strain temporal field that is consistent with the anatomical boundary and spatially continuous.

[0154] Based on the previously retained residual / confidence information and baseline alignment threshold, the initial surface strain time series field is quality controlled, i.e., low confidence frames (occlusion, reflection, fogging, film texture interference) are marked and removed, and missing frames are repaired with physically consistent temporal interpolation; then, the entire sequence is normalized in terms of dimensions / color levels and aligned with pressure timestamps to ensure that different stages can be compared laterally; finally, a surface displacement strain time series field that has passed quality control is obtained (used for threshold determination, compliance identification and safety envelope analysis), with frame-level quality labels attached.

[0155] In this embodiment, the refraction / distortion of the transparent airbag is eliminated and unified to the inherent coordinates of the wound edge through medium geometry reduction, avoiding systematic errors introduced by the vehicle and viewpoint. Subsequently, the non-pressurized components are stripped by correlation matching and optical flow fusion combined with baseline counterfactual stripping to obtain subpixel accuracy and temporally continuous causal displacement. Then, strain is reconstructed under the constraints of boundary conservation, radial compliance monotonicity, and topological preservation to suppress non-physical artifacts such as cracking / agglomeration and ensure consistency with anatomy and stress. Finally, residual / confidence and visual quality guards are used to complete the elimination and interpolation, dimensional and temporal alignment, and output a surface displacement-strain temporal field that can be directly used for threshold determination, compliance fitting, and safe envelope generation. This can significantly improve measurement accuracy, robustness, and repeatability, and reduce the risk of misjudgment in clinical intervention.

[0156] In an exemplary embodiment, a dual-stream counterfactual reduction process is performed on the primacy coordinate frame sequence to obtain a sub-pixel displacement temporal field, including steps 902 to 910. Wherein:

[0157] Step 902: Perform in-phase processing on the inherent coordinate frame sequence of the wound edge and the baseline imaging data to obtain the time frame pair sequence and the wound edge temporal index; Step 904: Perform dual-stream extraction on the correlation features and optical flow features in the time frame pair sequence to obtain the correlation displacement flow, optical flow displacement flow, and consistency residual map; Step 906: Based on the creation timing index and the consistency residual map, perform causal stripping on the relevant displacement flow and optical displacement flow to obtain the pressure-induced displacement timing. Step 908: Perform co-distillation processing on the pressure-induced displacement time series and the consistency residual map to obtain the sub-pixel displacement candidate field; Step 910: Based on the prior constraints of the creation boundary and the radial compliance constraints, the candidate field of subpixel displacement is subjected to constraint projection processing to obtain the temporal field of subpixel displacement.

[0158] Among them, the in-phase processing is a step of fine-grained alignment of the inherent coordinate frame sequence of the creation edge with the baseline imaging according to the timestamp, pressure record and phase correlation. Through resampling and phase correction, the acquisition rhythm difference and trigger jitter are eliminated, so that the observation frame at any time can be paired with the corresponding baseline frame.

[0159] The time frame sequence is a set of paired images arranged in chronological order after in-phase conversion. Each pair consists of "current frame - corresponding baseline or neighboring frame" and carries a unified timestamp and stress phase information.

[0160] Among them, the Creation Time Series Index is an index entry created for each time frame pair, recording metadata such as the pressure stage, step window, frame number range, and ROI mask, which facilitates rapid backtracking and cross-step reference during causal stripping and event localization.

[0161] Among them, the relevant features are texture blocks / descriptors oriented towards digital image correlation (DIC) and their matching responses (such as cross-correlation peaks, confidence, and sub-window interpolation shifts), which are used to measure the local block correspondence and shift estimation between frame pairs.

[0162] Among them, optical flow features are pixel-level motion information (such as pyramid optical flow vectors, gradient intensity, residuals and visibility markers) calculated based on brightness consistency and gradient constraints, which are used to characterize the fine-grained displacement field between consecutive frames.

[0163] Among them, dual-stream extraction involves simultaneously extracting "correlation features" and "optical flow features" for each time frame pair, calculating two displacement estimates separately, and combining them with local texture / confidence to generate a consistency metric.

[0164] Among them, the relevant displacement flow is a displacement vector field obtained by block matching and subpixel interpolation of relevant features. It is good at handling regions with clear textures and comes with matching confidence and occlusion hints.

[0165] Among them, optical flow displacement flow is a displacement vector field obtained by solving optical flow features through pyramid multi-scale solution. It is more sensitive to small continuous motion and weak texture regions, and comes with residual and stability indicators.

[0166] Among them, the consistency residual map is a difference intensity map (including threshold / confidence information) obtained by comparing the relevant displacement flow and the optical flow displacement flow pixel by pixel. It is used to mark the areas where the two estimates are inconsistent or affected by occlusion / reflection, and guide the weight allocation of subsequent stripping and fusion.

[0167] Among them, the causal stripping process uses the pressure stage marked by the creation time sequence index as the time reference, and combines it with the "no-pressure prediction trajectory" generated by the baseline frame to remove non-pressurized components such as respiration, body movement, and camera drift from the two displacement flows, retaining only the causal motion components caused by pressure changes.

[0168] Among them, the pressure-induced displacement time series is a sequence of displacement vector fields arranged over time after causal separation, representing the real tissue displacement "generated solely by pressure".

[0169] Among them, the co-distillation process uses the consistency residual map for gating, and merges and refines the relevant displacement flow and optical displacement flow according to spatial adaptive weights and time filtering (such as smoothing / Kalman update), suppressing single path deviation and forming candidate displacements with lower noise and better coherence.

[0170] Among them, the subpixel displacement candidate field is a high-precision, spatiotemporally continuous displacement vector candidate result output by the co-distillation process, which has been interpolated and smoothed at the subpixel level.

[0171] Among them, the a priori constraints of the wound edge boundary are normal / tangential behavior constraints (such as boundary normal continuity, non-outward rotation, and tangential smoothness) established based on the anatomical boundary of the wound edge and imaging geometry, which are used to ensure the physical consistency of the displacement field at the boundary.

[0172] Among them, radial compliance constraint decomposes the displacement into radial / tangential components with the creation edge as a reference, and applies monotonically non-reducing, amplitude limiting and priority requirements to the radial components to conform to the mechanical properties of "inward compliance" of the tissue under pressure.

[0173] Among them, the constraint projection process projects the subpixel displacement candidate field onto a feasible manifold that satisfies the boundary prior and radial compliance constraints, and back-projects the vectors that violate the constraints with the minimum change until the entire field achieves physical-topological consistency.

[0174] Specifically, the inherent coordinate frame sequence of the creation edge and the baseline imaging data are synchronously resampled according to the timestamp and pressure record, and the observation at each moment and its baseline reference are fine-grained phase corrected by the phase correlation / cross-correlation peak. After alignment, a one-to-one corresponding frame pair sequence is generated as a time frame pair sequence, and the creation edge time sequence index recording "pressure stage - time window - frame number" is output.

[0175] 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 two displacements are subtracted at the pixel level and combined with the local texture confidence to calculate the consistency residual map. At the same time, the confidence weights and occlusion cues of the two paths are retained to form the correlation displacement flow and optical flow displacement flow.

[0176] The pressure stage of each frame is determined based on the creation timing index, and the non-pressurized components such as background / breathing are stripped from the relevant displacement flow and optical flow displacement flow using the pressureless prediction trajectory generated by the baseline reference frame. Then, the low consistency region is masked by gating with the consistency residual map, and the residual displacement is constrained by the temporal monotonicity and physical direction (radial priority). The output only retains the timing of the pressure-induced displacement "born from pressure".

[0177] The information in the temporal sequence and consistency residual map of pressure-induced displacement is dynamically weighted according to spatial adaptive weights; combined with exponential smoothing and Kalman update in time to suppress noise drift, and refined at the sub-pixel level by spline / Lagrange interpolation, a spatiotemporally continuous and locally smooth sub-pixel displacement candidate field is obtained.

[0178] Based on the boundary prior constraints and radial compliance constraints, the subpixel displacement candidate field is projected onto the physically feasible manifold. During the projection process, the radial component is forced to be monotonically non-decreasing, and the tangential component is subjected to smoothing / divergence suppression. Normal continuity and non-outward turning constraints are applied at the boundary, and vectors that violate the constraints are back-projected with minimum change. After frame-by-frame iteration until convergence, the subpixel displacement temporal field that satisfies boundary conservation and radial compliance is output.

[0179] In this embodiment, in-phase conversion first eliminates acquisition rhythm differences and trigger jitter, avoiding spurious displacements caused by time mismatches. Dual-stream refining, with correlation / DIC and optical flow complementarity and unreliable regions marked by a consistent residual chart, improves the lower bound and interpretability of displacement estimation. Furthermore, causal stripping, under the constraint of the wound edge temporal index, removes non-pressurized components such as respiration / body motion, retaining only pressurized induced motion, significantly reducing systematic interference. Co-distillation utilizes residual gating for adaptive weighting and temporal filtering to obtain sub-pixel candidates with lower noise and better continuity. Finally, projection onto a physically feasible manifold under wound edge boundary prior and radial compliance constraints suppresses artifacts such as cracking / agglomeration and ensures boundary conservation. This significantly improves accuracy, robustness, and repeatability, and provides an auditable and directly callable causal displacement foundation for subsequent strain reconstruction and safety control.

[0180] In one exemplary embodiment, a deformation visualization monitoring system for the pressurization process of a wound rewarming chamber is provided, the system including a terminal and a computer device; The main core functions of computer equipment include: 1. Thermal Insulation Module: Utilizing ultra-thin flexible graphene material, it heats up in 3 seconds, evenly covering the wound surface with heat, avoiding localized burns or uneven temperature distribution caused by traditional hot compresses. Even if the outside temperature drops sharply, the temperature inside the chamber remains stable. Its outer layer is an aerospace-grade aerogel insulation layer, combined with vacuum sealing technology, effectively isolating cold air from the outside and ensuring continuous insulation for 6-12 hours in extremely cold environments.

[0181] 2. Control Module: The control panel design of the electronic blood pressure monitor can be improved. It features a built-in lithium battery (capable of 8 hours of continuous standby at -30℃), integrating Bluetooth and a mobile app into the control chip (with an embedded memory system for efficient heating and pressure control). A manual control interface is also available as a backup in case of emergencies. Charging is available via a UTC port, and solar charging is possible in extremely cold outdoor environments. A connection port (external backup device) is located at the bottom of the control module for connecting to the improved wrist blood pressure monitor, enabling data exchange via a data cable.

[0182] 3. Heating / pressurizing module: A TPE high transparency and flexible airbag and mesh heating wire are added to the outside of the insulation module. The airbag pressure is precisely controlled at 30-40 mmHg and the temperature at 36-38℃ through the control plate.

[0183] 4. Fixing Module: It is recommended to use a lightweight, heat-insulating, waterproof Cordura nylon + aerogel composite material that is heat-insulating, waterproof, wear-resistant, and easy to carry. High-strength, low-temperature resistant snap-on fasteners that can be quickly opened and closed with one hand are recommended for both ends. A 10x10cm window can be opened on the outer layer with sealed edges for easy observation of the wound.

[0184] 5. Vital Energy Monitoring Module: Utilizing a high-strength, cold-resistant composite webbing, the outer layer is made of abrasion-resistant and tear-resistant tactical-grade nylon, while the inner layer incorporates a flexible silicone anti-slip layer, providing a secure yet comfortable fit against the wrist. The webbing is equipped with self-adjusting buckles (made of cold-resistant rubber), maintaining resilience even in low temperatures and providing a reliable fixation in extreme environments, ensuring continuous and stable contact between the monitoring module and the skin for accurate collection of vital signs data. Heart rate and blood oxygen saturation monitoring employ advanced photoplethysmography (PPG) technology (based on changes in blood's absorption of light) to ensure accurate control of heart rate and blood oxygen saturation data within ±5% and ±2%, respectively. Blood pressure monitoring uses PPG (green LED + photodiode) to monitor changes in blood pressure volume through optical principles, fusing data with pressure sensor data, with measurement errors controlled within ±8mmHg. Respiratory rate is achieved through a fusion of PPG green LED illumination and an ECG sensor, capturing subtle fluctuations in vascular volume with respiration while simultaneously acquiring ECG signals to accurately calculate respiratory rate. The body temperature monitoring uses an infrared thermopile sensor, which can measure temperature non-contactly and avoid the risk of metal probes sticking to the skin in low-temperature environments.

[0185] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown. This computer device includes a processor, memory, input / output interfaces (I / O), and communication interfaces. Those skilled in the art will understand that... Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0186] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0187] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0188] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including 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 executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0189] 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 used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0190] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. 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 embodiments of the above methods.

[0191] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0192] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by 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