Image enhancement methods for wound assessment

Abstract

This invention relates to the field of photoelectric signal energy conditioning technology and discloses an image enhancement method for wound assessment, comprising: acquiring a multi-channel photoelectric raw signal stream under an unsteady excitation source; constructing a reference energy response vector in the spectrum and power coupling vector space; determining the energy overload characteristic value based on the vector displacement deviation of the signal and generating an energy distribution coefficient through reverse negative feedback logic; and dynamically modulating the power amplitude component according to the coefficient. Under the constraint of locking the characteristic energy of the spectrum component, this invention achieves precise suppression of transient amplitude surges and solves the technical problems of signal-to-noise ratio degradation and source-end bandwidth impedance characteristic truncation under unsteady conditions.

Description

Technical Field

[0001] This invention relates to the field of photoelectric signal energy conditioning technology, and in particular to an image enhancement method for wound assessment. Background Technology

[0002] In current engineering practices for remote wound diagnosis and chronic wound management, non-contact image acquisition based on mobile terminals has become the main way to obtain wound tissue information. General-purpose optical sensors are used to collect the reflectance spectrum signal of the wound surface in uncontrolled lighting environments, and signal processing pipelines based on Retinex theory or histogram equalization are used to correct the uneven illumination in order to restore the intrinsic reflectance characteristics of human tissue. Existing technologies are generally based on the engineering assumption that the image signal is a linear superposition of low-frequency ambient light component and high-frequency object reflection component, and filtering or mapping algorithms are designed accordingly to separate the two.

[0003] However, as a non-Lambertian surface with a high specular reflectance coefficient, the optical response characteristics of moist wounds deviate from the aforementioned linear superposition assumption. Under the illumination of non-standard point light sources such as ward ceiling lights or flashlights, the specular reflection components generated by exudate, gel, or necrotic liquefaction on the wound surface will nonlinearly couple and superimpose with the diffuse reflection components of the tissue. This physical signal aliasing leads to a technical contradiction in existing linear enhancement algorithms when processing such images: they cannot simultaneously achieve brightness balance and feature fidelity. Existing technologies have the following three main technical limitations when processing images of such moist wounds: 1. Strong coupling between illumination components and pathological semantics leads to feature misjudgment. Traditional algorithms tend to uniformly regard high-brightness pixels as well-lit areas or white necrotic tissue. However, in moist wounds, high-brightness areas are often caused by signal saturation due to specular reflection. Existing global or local contrast stretching algorithms lack a mechanism to identify this physical cause, and are prone to misinterpreting reflective artifacts. 1. Mistakenly enhancing features of necrotic tissue, or erasing superficial granulation textures with diagnostic value when suppressing highlights, resulting in truncation of semantic information; 2. The inherent contradiction between signal gain adjustment and noise suppression: wound images usually contain a high dynamic range of brightness distribution, including highly reflective liquid surfaces and deep cavities in shadow. Existing algorithms often use a single gain coefficient when improving the visibility of dark details, which inevitably amplifies the dark current noise of the sensor exponentially, resulting in severe salt-and-pepper noise interference in deep cavity areas, masking the true tissue topology; 3. The asynchronicity between color fidelity and brightness adjustment: conventional grayscale transformation mainly acts on the brightness channel, ignoring the intrinsic correlation between chroma and brightness in biological tissues under different pathological states. For example, when increasing overall brightness, deeply congested inflammatory tissue often presents a healthy pink appearance due to reduced saturation. This drift in color information directly weakens the image's ability to warn of infection risks.

[0004] To address the aforementioned issues, industry attempts have focused on introducing polarized light hardware to filter out reflections or utilizing deep learning networks for repair. However, the former increases the hardware cost and operational complexity of the system, while the latter is limited by the scarcity of medical annotation datasets and the inference latency of edge computing devices, making large-scale application in primary healthcare scenarios difficult. Therefore, the technical problem this invention aims to solve is how to construct a signal decoupling and feature reshaping mechanism that conforms to physical optics priors, using only single-frame color image data without introducing additional optical hardware, to achieve accurate separation and semantic restoration of ambient light interference, specular reflection noise, and intrinsic tissue texture in wet wound images. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an image enhancement method for wound assessment, comprising the following steps:

[0006] The multi-channel raw photoelectric signal stream is acquired by the photoelectric conversion interface under unsteady excitation source conditions. The raw photoelectric signal stream contains spectral component signals and power amplitude component signals that characterize the impedance characteristics of the source load.

[0007] Within the preset spectrum-power coupling vector space, a reference vector characterizing the reference power response mode of the region is constructed based on the local power density statistical characteristics of the original photoelectric signal flow.

[0008] Perform a signal energy anomaly feature extraction step based on vector deviation, calculate the vector displacement deviation of the signal unit to be conditioned relative to the reference vector in the spectrum-power coupling vector space, and confirm the vector displacement deviation as the energy overload characteristic value characterizing the transient amplitude surge intensity;

[0009] A nonlinear power gain control command is generated for the original photoelectric signal stream. The nonlinear power gain control command is based on the energy overload characteristic value and is calculated by reverse negative feedback suppression logic to obtain the power distribution coefficient used to correct the local signal power density.

[0010] The nonlinear power gain control command is executed, and the power amplitude component signal in the original photoelectric signal stream is dynamically modulated according to the power distribution coefficient. Under the constraint that the source-end frequency band impedance characteristic energy carried by the locked spectral component signal is not attenuated, the conditioned signal data stream with amplitude surge suppression and signal-to-noise ratio optimization is output.

[0011] Preferably, the step of constructing a reference vector characterizing the reference power response mode of the region specifically includes: defining a sliding sampling window centered on the signal unit to be conditioned; identifying the power amplitude level of all signal units within the sliding sampling window; filtering out oversaturated signal units whose power amplitude level exceeds the cutoff threshold of a preset linear response interval, and retaining effective signal units in the linear operating region; calculating the mean of the spectral components and the mean of the power amplitude components of all effective signal units in the spectrum-power coupling vector space; and confirming the two-dimensional vector composed of the mean of the spectral components and the mean of the power amplitude components as the reference vector.

[0012] Preferably, the step of extracting signal energy anomaly features based on vector deviation specifically includes: mapping the signal unit to be conditioned to the spectrum-power coupling vector space to determine its actual state coordinate position; establishing a differential vector pointing from the endpoint of the reference vector to the actual state coordinate position; calculating the magnitude of the differential vector and defining the magnitude as the vector displacement deviation; constructing a continuously distributed impedance mismatch weighted field, mapping the vector displacement deviation to a scalar weight value in the impedance mismatch weighted field, whereby the scalar weight value characterizes the energy confidence of the signal unit under nonlinear impedance mismatch interference.

[0013] Preferably, the step of generating a nonlinear power gain control command for the original photoelectric signal stream specifically includes: introducing a logic-gated variable based on the saturation level of the spectral channel; real-time detection of the saturation component level of the signal unit to be conditioned; activating a nonlinear attenuation control loop based on the logic-gated variable when the vector displacement deviation exceeds a preset interference tolerance threshold; and calculating the power distribution coefficient using the nonlinear attenuation control loop according to the following transfer function relationship: ,in, The output power distribution coefficient. The system's fundamental gain constant is the preset value, and ΔV is the vector displacement deviation. α is the preset maximum allowable deviation amplitude, α is the response sensitivity adjustment factor, and S is the saturation component level of the signal unit to be conditioned.

[0014] Preferably, the step of executing the nonlinear power gain control command follows the cross-channel asynchronous coupling processing logic, specifically including: separating the power transmission channel and the spectrum modulation channel of the original optoelectronic signal stream; applying a power distribution coefficient only to the power transmission channel to implement negative feedback adjustment of the power amplitude component signal level; locking the transmission gain response ratio of the spectrum modulation channel to a preset reference value to maintain the frequency domain response linearity of the spectrum component signal during dynamic power modulation; and performing orthogonal vector synthesis of the adjusted power amplitude component signal and the original spectrum component signal to generate a conditioned signal data stream.

[0015] Preferably, the method for solving the signal truncation problem caused by unsteady point source excitation and generating a nonlinear power gain control command for the original photoelectric signal stream further includes: identifying a high-saturation frequency band region in the original photoelectric signal stream corresponding to a specific low-impedance source characteristic; constructing a reverse gain protection model and setting a characteristic protection damping interval for the high-saturation frequency band region; when calculating the power distribution coefficient, if the parameters of the signal unit to be conditioned fall into the characteristic protection damping interval, the power suppression slope for the signal unit is reduced to ensure that the high-energy spectral characteristics corresponding to a specific source impedance maintain their separation in the signal energy spectrum during power adjustment.

[0016] Preferably, the method also involves signal energy compensation in the low potential dead zone, specifically including: detecting the DC base component level in the original photoelectric signal stream; when the DC base component level is lower than a preset minimum potential threshold, generating an independent low potential compensation gain command; linearly weighting the low potential compensation gain command and the nonlinear power gain control command; while increasing the signal energy density in the low potential region, using the vector displacement deviation as a feedback damping factor to limit the energy amplification of the accompanying additive noise.

[0017] Preferably, before the step of acquiring the multi-channel photoelectric raw signal stream collected by the photoelectric conversion interface under non-steady-state excitation source conditions, the method further includes: initializing the preamplifier gain parameters of the photoelectric conversion interface; performing a full-band pre-scan operation to acquire the spectral distribution data of the environmental excitation source; establishing a spectral drift correction matrix for a specific excitation source type based on the spectral distribution data; and using the spectral drift correction matrix to perform reference zero-point calibration on the subsequently acquired photoelectric raw signal stream.

[0018] Preferably, the method is applied to a remote signal processing terminal with limited energy constraints. The step of executing nonlinear power gain control commands further includes: real-time monitoring of the power supply unit voltage status of the remote signal processing terminal; when the power supply unit voltage status is lower than a preset low-voltage protection threshold, automatically switching to a low-power operation mode; in the low-power operation mode, reducing the sampling discretization frequency of the spectrum-power coupling vector space, and using a piecewise linear fitting function to replace the high-order exponential operation in the reverse negative feedback suppression logic, so as to reduce the dynamic switching power consumption of the digital signal processing unit.

[0019] Preferably, after outputting the conditioned signal data stream that has completed amplitude surge suppression and signal-to-noise ratio optimization, the method further includes: converting the conditioned signal data stream into a visual display driving voltage signal; driving the display execution unit to present the source load distribution map; and simultaneously outputting a signal link quality assessment data packet containing the statistical distribution of energy overload characteristic values, and encapsulating the data packet through a standard data interface protocol for transmission to an external source status trend analysis module.

[0020] The beneficial effects of this invention are as follows:

[0021] 1. By constructing a local statistical model based on the relationship between chromaticity and luminance vectors, this method addresses the problem of pathological feature truncation caused by specular reflection in images of wet wounds. Instead of simply applying uniform grayscale suppression to bright areas, this method calculates the coupling deviation of pixels relative to the statistical center of neighboring normal tissue in the color space, thereby quantifying and identifying areas where physical signals are obscured by reflection. The system uses this coupling deviation as a weighting operator to guide the texture features of uncontaminated tissue in the neighborhood to penetrate into the bright areas with weighted intensity. This feature reshaping mechanism based on physical priors allows areas that originally lost information due to light saturation to be filled with statistically significant texture details according to the continuity of surrounding biological tissues. This eliminates highlight artifacts while ensuring the semantic continuity of granulation tissue or necrotic tissue in the eyes of the image segmentation algorithm, avoiding misjudgment of areas due to signal loss.

[0022] 2. By utilizing the dual properties of illumination components in the frequency domain, this method overcomes the technical drawback of traditional Retinex-type algorithms that simultaneously erase the three-dimensional features of the tissue surface when removing ambient light. During the illumination decoupling process, this method identifies that the high spatial frequency fluctuations remaining in the illumination component image actually correspond to the tiny projections generated by granulation particles or wound edge wrinkles, rather than ambient light. The system extracts this high-frequency residual through high-pass filtering and uses it as a modulation signal to inject back into the reflectivity image after the illumination has been removed. This signal processing flow mathematically achieves the precise separation and reuse of the macroscopic ambient light gradient and the surface structure light gradient, enabling the enhanced two-dimensional image to re-present the granularity and depth information reflecting tissue activity under flat lighting conditions. This provides input data containing quasi-three-dimensional morphological features for subsequent automatic evaluation algorithms based on texture roughness.

[0023] 3. An adaptive gain clamping mechanism based on illumination intensity is established to address the signal-to-noise ratio (SNR) degradation problem encountered during enhancement of deep cavities or sinus tracts. The system repeatedly uses the calculated ambient illumination component as a confidence index of the regional SNR and constructs a maximum gain constraint function positively correlated with illumination intensity. When performing nonlinear stretching on the image, this function forcibly constrains the contrast enhancement amplitude in low-illuminance areas. This negative feedback control logic ensures that in physically dark areas with insufficient photons, the algorithm will not exponentially amplify the dark current noise of the sensor in pursuit of brightness balance. This allows deep cavities to maintain visual depth and purity in the enhanced image, effectively preventing the risk of misidentifying high-frequency noise as necrotic tissue texture and ensuring the correct representation of wound topological depth. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein:

[0025] Figure 1 This is a flowchart of the steps of the image enhancement method for wound assessment according to the present invention;

[0026] Figure 2 This is a diagram illustrating the principle elements and signal processing logic architecture of the image enhancement method of the present invention. Detailed Implementation

[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0028] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0029] Secondly, an embodiment or embodiment referred to herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. An embodiment appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0030] This invention is described in detail with reference to the schematic diagrams. When describing the embodiments of this invention, for ease of explanation, the cross-sectional views of the device structure will be partially enlarged without adhering to the general scale. Moreover, the schematic diagrams are only examples and should not limit the scope of protection of this invention. In addition, in actual manufacturing, the three-dimensional spatial dimensions of length, width and depth should be included.

[0031] Furthermore, in the description of this invention, it should be noted that the terms such as "upper," "lower," "inner," and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or component referred to has a specific orientation, or is constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0032] Unless otherwise explicitly specified and limited, the terms installation, connection, and linking in this invention should be interpreted broadly. For example, they can refer to fixed connection, detachable connection, or integrated connection; similarly, they can refer to mechanical connection, electrical connection, or direct connection, or indirect connection through an intermediate medium, or internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0033] An image enhancement method for wound assessment includes the following steps:

[0034] The multi-channel raw photoelectric signal stream is acquired by the photoelectric conversion interface under unsteady excitation source conditions. The raw photoelectric signal stream contains spectral component signals and power amplitude component signals that characterize the impedance characteristics of the source load.

[0035] Within the preset spectrum-power coupling vector space, a reference vector characterizing the reference power response mode of the region is constructed based on the local power density statistical characteristics of the original photoelectric signal flow.

[0036] Perform a signal energy anomaly feature extraction step based on vector deviation, calculate the vector displacement deviation of the signal unit to be conditioned relative to the reference vector in the spectrum-power coupling vector space, and confirm the vector displacement deviation as the energy overload characteristic value characterizing the transient amplitude surge intensity;

[0037] A nonlinear power gain control command is generated for the original photoelectric signal stream. The nonlinear power gain control command is based on the energy overload characteristic value and is calculated by reverse negative feedback suppression logic to obtain the power distribution coefficient used to correct the local signal power density.

[0038] The nonlinear power gain control command is executed, and the power amplitude component signal in the original photoelectric signal stream is dynamically modulated according to the power distribution coefficient. Under the constraint that the source-end frequency band impedance characteristic energy carried by the locked spectral component signal is not attenuated, the conditioned signal data stream with amplitude surge suppression and signal-to-noise ratio optimization is output.

[0039] Preferably, the step of constructing a reference vector characterizing the reference power response mode of the region specifically includes: defining a sliding sampling window centered on the signal unit to be conditioned; identifying the power amplitude level of all signal units within the sliding sampling window; filtering out oversaturated signal units whose power amplitude level exceeds the cutoff threshold of a preset linear response interval, and retaining effective signal units in the linear operating region; calculating the mean of the spectral components and the mean of the power amplitude components of all effective signal units in the spectrum-power coupling vector space; and confirming the two-dimensional vector composed of the mean of the spectral components and the mean of the power amplitude components as the reference vector.

[0040] Preferably, the step of extracting signal energy anomaly features based on vector deviation specifically includes: mapping the signal unit to be conditioned to the spectrum-power coupling vector space to determine its actual state coordinate position; establishing a differential vector pointing from the endpoint of the reference vector to the actual state coordinate position; calculating the magnitude of the differential vector and defining the magnitude as the vector displacement deviation; constructing a continuously distributed impedance mismatch weighted field, mapping the vector displacement deviation to a scalar weight value in the impedance mismatch weighted field, whereby the scalar weight value characterizes the energy confidence of the signal unit under nonlinear impedance mismatch interference.

[0041] Preferably, the step of generating a nonlinear power gain control command for the original photoelectric signal stream specifically includes: introducing a logic-gated variable based on the saturation level of the spectral channel; real-time detection of the saturation component level of the signal unit to be conditioned; activating a nonlinear attenuation control loop based on the logic-gated variable when the vector displacement deviation exceeds a preset interference tolerance threshold; and calculating the power distribution coefficient using the nonlinear attenuation control loop according to the following transfer function relationship: ,in, The output power distribution coefficient. The system's fundamental gain constant is the preset value, and ΔV is the vector displacement deviation. α is the preset maximum allowable deviation amplitude, α is the response sensitivity adjustment factor, and S is the saturation component level of the signal unit to be conditioned.

[0042] Preferably, the step of executing the nonlinear power gain control command follows the cross-channel asynchronous coupling processing logic, specifically including: separating the power transmission channel and the spectrum modulation channel of the original optoelectronic signal stream; applying a power distribution coefficient only to the power transmission channel to implement negative feedback adjustment of the power amplitude component signal level; locking the transmission gain response ratio of the spectrum modulation channel to a preset reference value to maintain the frequency domain response linearity of the spectrum component signal during dynamic power modulation; and performing orthogonal vector synthesis of the adjusted power amplitude component signal and the original spectrum component signal to generate a conditioned signal data stream.

[0043] Preferably, the method for solving the signal truncation problem caused by unsteady point source excitation and generating a nonlinear power gain control command for the original photoelectric signal stream further includes: identifying a high-saturation frequency band region in the original photoelectric signal stream corresponding to a specific low-impedance source characteristic; constructing a reverse gain protection model and setting a characteristic protection damping interval for the high-saturation frequency band region; when calculating the power distribution coefficient, if the parameters of the signal unit to be conditioned fall into the characteristic protection damping interval, the power suppression slope for the signal unit is reduced to ensure that the high-energy spectral characteristics corresponding to a specific source impedance maintain their separation in the signal energy spectrum during power adjustment.

[0044] Preferably, the method also involves signal energy compensation in the low potential dead zone, specifically including: detecting the DC base component level in the original photoelectric signal stream; when the DC base component level is lower than a preset minimum potential threshold, generating an independent low potential compensation gain command; linearly weighting the low potential compensation gain command and the nonlinear power gain control command; while increasing the signal energy density in the low potential region, using the vector displacement deviation as a feedback damping factor to limit the energy amplification of the accompanying additive noise.

[0045] Preferably, before the step of acquiring the multi-channel photoelectric raw signal stream collected by the photoelectric conversion interface under non-steady-state excitation source conditions, the method further includes: initializing the preamplifier gain parameters of the photoelectric conversion interface; performing a full-band pre-scan operation to acquire the spectral distribution data of the environmental excitation source; establishing a spectral drift correction matrix for a specific excitation source type based on the spectral distribution data; and using the spectral drift correction matrix to perform reference zero-point calibration on the subsequently acquired photoelectric raw signal stream.

[0046] Preferably, the method is applied to a remote signal processing terminal with limited energy constraints. The step of executing nonlinear power gain control commands further includes: real-time monitoring of the power supply unit voltage status of the remote signal processing terminal; when the power supply unit voltage status is lower than a preset low-voltage protection threshold, automatically switching to a low-power operation mode; in the low-power operation mode, reducing the sampling discretization frequency of the spectrum-power coupling vector space, and using a piecewise linear fitting function to replace the high-order exponential operation in the reverse negative feedback suppression logic, so as to reduce the dynamic switching power consumption of the digital signal processing unit.

[0047] Preferably, after outputting the conditioned signal data stream that has completed amplitude surge suppression and signal-to-noise ratio optimization, the method further includes: converting the conditioned signal data stream into a visual display driving voltage signal; driving the display execution unit to present the source load distribution map; and simultaneously outputting a signal link quality assessment data packet containing the statistical distribution of energy overload characteristic values, and encapsulating the data packet through a standard data interface protocol for transmission to an external source status trend analysis module.

[0048] Example 1: In a remote terminal processing environment for processing non-steady-state photoelectric signals, when the photoelectric conversion interface is under uncontrolled point light source excitation, the acquired multi-channel raw photoelectric signal stream not only contains effective spectral components characterizing the intrinsic impedance of biological tissue, but also superimposed high-energy power amplitude components generated by specular reflection from the liquefied area of ​​the wound. Nonlinear signal coupling leads to local signal-to-noise ratio degradation, making it difficult for conventional linear filtering algorithms to suppress highlights while preserving the underlying texture features. To address the energy anomaly problem in the signal link, the digital signal processing unit first constructs a reference energy response vector in the spectral and power coupling vector space. The specific process includes defining a sliding sampling window centered on the current signal unit to be conditioned and covering 15×15 pixel units, traversing the power amplitude level of all signal units within the window, filtering out nonlinear signal points in an oversaturated state using a preset linear response interval cutoff threshold, and retaining only effective signal units in the linear operating region; the system calculates the mean of the spectral components and the mean of the power amplitude components of the effective signal units in the coupling space, and synthesizes these two mean components into a two-dimensional vector, which is identified as the reference vector characterizing the reference power response mode of the local region, thereby establishing a dynamic drift energy reference surface that can adapt to the fluctuations of unsteady excitation sources.

[0049] Based on this dynamic benchmark, the system performs a signal energy anomaly feature extraction step based on vector deviation, mapping the current signal unit to be conditioned onto the spectrum and power coupled vector space to determine its actual state coordinate position, and establishing a differential vector pointing from the endpoint of the reference vector to this actual state coordinate position; the magnitude of this differential vector is calculated and defined as the vector displacement deviation. This physical quantity directly quantifies the energy escape degree of the current signal unit relative to the local reference mode and is identified as the energy overload characteristic value characterizing the transient amplitude surge intensity. Subsequently, the system generates a nonlinear power gain control command for the original photoelectric signal stream. This command introduces reverse negative feedback suppression logic and uses the calculated energy overload characteristic value to correct the local power density. During the calculation process, the system detects the saturation component level value of the signal unit to be conditioned in real time. And introduce configuration as Response sensitivity adjustment factor Based on the transfer function relationship Calculate the power distribution coefficient ,in: The output power distribution coefficient. The preset system fundamental gain constant, The maximum permissible deviation amplitude is preset; finally, the system executes nonlinear power gain control commands according to the cross-channel asynchronous coupling processing logic, specifically including: separating the original photoelectric signal stream into a power transmission channel and a spectrum modulation channel; and applying a power distribution coefficient only to the power transmission channel. The power amplitude component signal level is adjusted by negative feedback; the transmission gain response ratio of the spectrum modulation channel is locked to a preset reference value; the processing unit performs orthogonal vector synthesis of the adjusted power amplitude component signal and the original spectrum component signal to generate a conditioned signal data stream. This data stream removes specular reflection interference while retaining the color and texture details that reflect the material properties of the source.

[0050] Example 2: In a verification test scenario for dynamic nonlinear distortion correction of photoelectric signals, the experiment aimed to quantitatively evaluate the ability of the nonlinear power gain control strategy of this invention to suppress high-saturation amplitude surges while preserving low-amplitude frequency band details. This experiment constructed a test platform comprising a multilayer biomimetic phantom simulating the optical properties of biological tissue and a high-frequency dynamic light source modulation system. The multilayer biomimetic phantom was formed by mixing and solidifying a specific concentration of lipid emulsion with hemoglobin powder, and its surface was covered with a 0.5 mm thick saline film to simulate the specular reflection characteristics of the liquefied area of ​​a wound. The light source modulation system was configured to output pulsed light with a randomly varying frequency between 10 Hz and 50 Hz to simulate unstable... Under point source excitation conditions, the data acquisition end uses a photoelectric conversion interface with 16-bit quantization depth and 1000fps sampling rate to acquire the original photoelectric signal stream containing the power transmission channel and the spectrum modulation channel. To verify the stability of this scheme in a real engineering environment, Gaussian white noise with a signal-to-noise ratio of 15dB was actively injected into the original signal link, and 50Hz power frequency interference with an amplitude of 10% of the signal peak value was superimposed to simulate the electromagnetic environment disturbance in the clinical setting. Two parallel control groups were set up in the experiment: the control group used the conventional linear filtering algorithm (CLAHE) to process the signal; the sample group of this invention used the nonlinear power gain control method based on the spectrum-power coupling vector space in this specification.

[0051] After the experiment started, the data acquisition system captured a raw signal stream lasting 5 seconds. For the sample group of this invention, the digital signal processing unit performed real-time construction of the reference energy response vector. Within a 15×15 pixel sliding window, the system identified and eliminated oversaturation points exceeding the cutoff threshold of the linear response interval (set to 85% of the sensor's full scale). The system calculated the vector displacement deviation ΔV of each effective signal unit. The data showed that in the high-gloss area of ​​the simulated specular reflection, the ΔV value exhibited a transient peak, with an amplitude exceeding three times the local reference mean. This was determined by the system to be an energy overload characteristic value. Subsequently, the system based on the transfer function... This generates a nonlinear power gain control command, which is used in this experiment. Set to 1.0. The sensor's dynamic range was set to 50%, and the response sensitivity adjustment factor α was set to 2.5. Process monitoring data showed that as ΔV increased, the output power distribution coefficient... It exhibits a non-linear exponential decay trend, especially when ΔV approaches... Within the interval, It rapidly drops below 0.1, achieving strong suppression of high-energy surges; while in the low ΔV texture region, Maintaining a value above 0.9 ensures lossless transmission of underlying details. In stark contrast, while the control group suppressed brightness in highlight areas, its histogram showed grayscale discontinuities, resulting in the loss of underlying texture information. The final signal quality assessment data confirmed the superiority of this solution. After processing, the peak signal-to-noise ratio (PSNR) of the output signal of the present invention sample increased from the original 18.5dB to 28.2dB. Furthermore, the chromaticity vector error (CVE) of the present invention sample was recorded at 1.2%, far lower than the 4.5% of the control group. By implementing negative feedback regulation in the power transmission channel and locking the gain response ratio of the spectral modulation channel, the system suppressed specular reflection interference while retaining the spectral fingerprint information characterizing biological tissue properties. This verifies the engineering effectiveness of the cross-channel asynchronous coupling processing logic in solving the problems of signal-to-noise ratio degradation and feature truncation under non-steady-state excitation.

[0052] Example 3: Before the photoelectric imaging system performs the formal wound signal acquisition task, the central processing unit forcibly initiates an adaptive parameter calibration procedure based on the sensor's physical characteristics to establish the geometric reference of the spectrum and power coupling vector space and the physical boundary of the nonlinear control parameters, define the coordinate system of the coupling vector space, and select the normalized power amplitude output from the photoelectric conversion interface. As the vertical axis, its value range is locked between 0 and 1, and the Shannon entropy of the multi-channel spectral signal is used. The horizontal axis is used to characterize the frequency domain complexity of the signal, thus constructing a two-dimensional orthogonal feature space. In this space, the state of any signal unit to be conditioned is uniquely mapped to a two-dimensional coordinate point. This eliminates the ambiguity in vector calculations caused by differences in physical dimensions; based on this spatial definition, the system performs linearity stress tests on the dynamic response range of the photoelectric conversion interface to determine key parameters. To determine the physical truth value, the test module injects a standard white light excitation signal with stepwise increasing intensity into the sensor and simultaneously monitors the power amplitude response of the output signal. The system calculates the first derivative of the response curve and defines the voltage offset corresponding to the derivative decreasing to 95% of the linear response reference value as... This step ensures It is no longer a fixed empirical constant, but a dynamic boundary value that is strictly bound to the physical saturation characteristics of the current hardware device.

[0053] Next, the system enters the automatic optimization calculation process for the response sensitivity adjustment factor α. Within the nonlinear cutoff region above the linear saturation critical point, the processing unit selects at least five discrete oversaturated response sampling points and extracts the actual vector magnitudes of these sampling points in the coupled vector space. The system constructs an exponential decay model with α as an undetermined coefficient, aiming to simulate an ideal anti-high-gloss human eye visual response curve. The transfer function is then solved iteratively using the least squares method. The root mean square error between the generated suppression curve and the ideal response curve is minimized. When the calculation converges, the system locks the α value obtained in the current iteration as a fixed parameter for subsequent real-time processing. This process transforms the setting logic of α from subjective selection to mathematical fitting result based on the minimum error criterion, ensuring that the system can obtain the optimal nonlinear suppression effect under different lighting conditions, and completing the logical closed loop from hardware physical properties to algorithm control parameters.

[0054] Example 4: To address the spatial inconsistency and spectral response differences in the photoelectric conversion channel, the system enforces a full-field luminous flux calibration procedure based on a Lambertian diffuse reflection reference surface before factory deployment. The modulated light source is controlled to output maximum power and project onto a standard grayscale plate with a constant reflectivity of 18%. The sensor array acquires the corresponding luminous flux distribution map and calculates the gain compensation coefficient matrix for each pixel relative to the mean response value at the center of the field of view. The matrix is ​​stored in non-volatile memory and performs point-to-point multiplication with the original signal in the front-end processing link of each signal acquisition cycle. This eliminates fixed-mode noise introduced by lens vignetting effect and photosensitive component manufacturing tolerance at the physical level, ensuring that the consistency of subsequent spectral and power coupling vector space measurements depends only on the biophysical properties of the wound being measured.

[0055] To address the thermal noise baseline drift caused by increased junction temperature during long-term equipment operation, the system incorporates dynamic dark current clamping and zero-point reset maintenance logic. This logic is automatically activated during the transient interval of each pulse excitation sequence, updating the dark current drift tensor by extinguishing the modulation light source and capturing the sensor's background noise signal with the highest sampling sensitivity. In the signal preprocessing stage, this tensor is deducted from the original photoelectric signal stream in real time to establish a zero-level reference that adapts to the ambient temperature and device aging state. This ensures that even under harsh conditions where the signal-to-noise ratio is squeezed by thermal noise, the vector displacement deviation ΔV in the transfer function can still accurately characterize the effective energy overload caused by external light intensity surges rather than thermal fluctuation interference from internal circuits.

[0056] Example 5: To ensure consistent system deployment and model adaptability across diverse engineering sites, this example discloses a standardized offline parameter calibration and system initialization procedure, which is mandatory during equipment installation and commissioning. The system performs on-site calibration of the spectral response baseline. Engineers need to place a standard whiteboard with a constant reflectivity of 99% at the target wound location. The system controls the modulation light source to output a full-power pulse sequence, collects the saturation response values ​​of each spectral channel, and calculates the deviation ratio between the saturation response values ​​and the factory-preset reference values. If any channel If the deviation exceeds ±5%, the system will automatically adjust the analog front-end gain register of that channel until the deviation returns to the tolerance range, thereby eliminating spectral distortion introduced by light source aging or optical path contamination.

[0057] The system enters an adaptive model parameter initialization process for specific wound types. Based on a pre-set pathological feature database, the system extracts the typical impedance distribution feature vector corresponding to the current wound type and injects it into the inverse gain protection model. The system performs Monte Carlo simulations to iteratively optimize the feature protection damping coefficient ζ within a preset parameter space. The objective function is set to minimize the signal truncation error in the high-saturation frequency band. When the rate of change of the objective function value for three consecutive iterations is less than... At this time, the system locks the current ζ value and writes it as the default operating parameter for this wound type into the protected sector of the configuration file; finally, the system performs online self-testing and fault-tolerant configuration of the entire link logic function. The processing unit sequentially activates the nonlinear attenuation control loop and the cross-channel asynchronous coupling logic, injects a set of digital test signals containing preset amplitude surges and spectral interference into the input, and monitors the power distribution coefficient at the output in real time. Compared to the signal-to-noise ratio of the conditioned signal, if within the test period, If the system fails to produce the expected attenuation response according to the transfer function, or if the output signal-to-noise ratio is lower than the preset minimum usable threshold, the system will trigger an abnormal alarm and automatically switch to the safety bypass mode, directly outputting the unconditioned raw signal to prevent the complete loss of diagnostic information due to algorithm logic errors, and to ensure the basic availability and safety of the system under extreme abnormal conditions.

[0058] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the protection scope of the present invention.

Claims

1. An image enhancement method for wound assessment, characterized in that, Includes the following steps: The multi-channel raw photoelectric signal stream is acquired by the photoelectric conversion interface under unsteady excitation source conditions. The raw photoelectric signal stream contains spectral component signals and power amplitude component signals that characterize the impedance characteristics of the source load. Within the preset spectrum-power coupling vector space, a reference vector characterizing the reference power response mode of the region is constructed based on the local power density statistical characteristics of the original photoelectric signal flow. Perform a signal energy anomaly feature extraction step based on vector deviation, calculate the vector displacement deviation of the signal unit to be conditioned relative to the reference vector in the spectrum-power coupling vector space, and confirm the vector displacement deviation as the energy overload characteristic value characterizing the transient amplitude surge intensity; A nonlinear power gain control command is generated for the original photoelectric signal stream. The nonlinear power gain control command is based on the energy overload characteristic value and is calculated by reverse negative feedback suppression logic to obtain the power distribution coefficient used to correct the local signal power density. The nonlinear power gain control command is executed, and the power amplitude component signal in the original photoelectric signal stream is dynamically modulated according to the power distribution coefficient. Under the constraint that the source-end frequency band impedance characteristic energy carried by the locked spectral component signal is not attenuated, the conditioned signal data stream with amplitude surge suppression and signal-to-noise ratio optimization is output.

2. The image enhancement method for wound assessment according to claim 1, characterized in that, The steps for constructing a reference vector characterizing the baseline power response mode of the region specifically include: defining a sliding sampling window centered on the signal unit to be conditioned; identifying the power amplitude level of all signal units within the sliding sampling window; filtering out oversaturated signal units whose power amplitude level exceeds the cutoff threshold of a preset linear response interval, and retaining effective signal units in the linear operating region; calculating the mean of the spectral components and the mean of the power amplitude components of all effective signal units in the spectrum-power coupling vector space; and confirming the two-dimensional vector composed of the mean of the spectral components and the mean of the power amplitude components as the reference vector.

3. The image enhancement method for wound assessment according to claim 1, characterized in that, The signal energy anomaly feature extraction step based on vector deviation is performed, specifically including: mapping the signal unit to be conditioned to the spectrum-power coupling vector space to determine its actual state coordinate position; establishing a differential vector pointing from the end point of the reference vector to the actual state coordinate position; calculating the magnitude of the differential vector and defining the magnitude as the vector displacement deviation; constructing a continuously distributed impedance mismatch weighted field, mapping the vector displacement deviation to a scalar weight value in the impedance mismatch weighted field, and the scalar weight value characterizing the energy confidence of the signal unit under nonlinear impedance mismatch interference.

4. The image enhancement method for wound assessment according to claim 3, characterized in that, The steps for generating nonlinear power gain control commands for the original photoelectric signal stream specifically include: introducing a logic-gated variable based on the saturation level of the spectral channel; real-time detection of the saturation component level of the signal unit to be conditioned; activating a nonlinear attenuation control loop based on the logic-gated variable when the vector displacement deviation exceeds a preset interference tolerance threshold; and calculating the power distribution coefficient using the nonlinear attenuation control loop according to the following transfer function relationship: ,in, The output power distribution coefficient. The system's fundamental gain constant is the preset value, and ΔV is the vector displacement deviation. α is the preset maximum allowable deviation amplitude, α is the response sensitivity adjustment factor, and S is the saturation component level of the signal unit to be conditioned.

5. The image enhancement method for wound assessment according to claim 1, characterized in that, The steps for executing nonlinear power gain control commands follow cross-channel asynchronous coupling processing logic, specifically including: separating the power transmission channel and the spectrum modulation channel of the original optoelectronic signal stream; applying a power distribution coefficient only to the power transmission channel to implement negative feedback adjustment of the power amplitude component signal level; locking the transmission gain response ratio of the spectrum modulation channel to a preset reference value to maintain the frequency domain response linearity of the spectrum component signal during dynamic power modulation; and performing orthogonal vector synthesis of the adjusted power amplitude component signal and the original spectrum component signal to generate a conditioned signal data stream.

6. The image enhancement method for wound assessment according to claim 1, characterized in that, The method is used to solve the signal truncation problem caused by unsteady point source excitation and generates a nonlinear power gain control command for the original photoelectric signal stream. It also includes the following steps: identifying a high-saturation frequency band region in the original photoelectric signal stream that corresponds to the characteristics of a specific low-impedance source; constructing a reverse gain protection model and setting a characteristic protection damping interval for the high-saturation frequency band region; and when calculating the power distribution coefficient, if the parameters of the signal unit to be conditioned fall into the characteristic protection damping interval, then the power suppression slope for that signal unit is reduced.

7. The image enhancement method for wound assessment according to claim 1, characterized in that, The method also involves signal energy compensation in low potential dead zones, specifically including: detecting the DC base component level in the original photoelectric signal stream; generating an independent low potential compensation gain command when the DC base component level is lower than a preset minimum potential threshold; linearly weighting the low potential compensation gain command and the nonlinear power gain control command; and while increasing the signal energy density in the low potential region, using the vector displacement deviation as a feedback damping factor to limit the energy amplification of associated additive noise.

8. The image enhancement method for wound assessment according to claim 1, characterized in that, Before acquiring the multi-channel raw photoelectric signal stream collected by the photoelectric conversion interface under unsteady excitation source conditions, the method further includes: initializing the preamplifier gain parameters of the photoelectric conversion interface; performing a full-band pre-scan operation to acquire the spectral distribution data of the environmental excitation source; establishing a spectral drift correction matrix for a specific excitation source type based on the spectral distribution data; and using the spectral drift correction matrix to perform reference zero-point calibration on the subsequently acquired raw photoelectric signal stream.

9. The image enhancement method for wound assessment according to claim 1, characterized in that, The method is applied to a remote signal processing terminal with limited energy constraints. The steps of executing nonlinear power gain control commands include: real-time monitoring of the power supply unit voltage status of the remote signal processing terminal; when the power supply unit voltage status is lower than a preset low-voltage protection threshold, automatically switching to a low-power operation mode; in the low-power operation mode, reducing the sampling discretization frequency of the spectrum-power coupling vector space, and using a piecewise linear fitting function to replace the high-order exponential operation in the reverse negative feedback suppression logic, so as to reduce the dynamic switching power consumption of the digital signal processing unit.

10. The image enhancement method for wound assessment according to claim 1, characterized in that, After the steps of outputting the conditioned signal data stream that has completed amplitude surge suppression and signal-to-noise ratio optimization are completed, the process also includes: converting the conditioned signal data stream into a visual display driving voltage signal; driving the display execution unit to present the source load distribution map; and simultaneously outputting a signal link quality assessment data packet containing the statistical distribution of energy overload characteristic values, and encapsulating the data packet through a standard data interface protocol for transmission to an external source status trend analysis module.

Images