Search

CN-122004853-A - Blood oxygen distribution imaging method and system based on double-band diffuse reflection space mapping

CN122004853ACN 122004853 ACN122004853 ACN 122004853ACN-122004853-A

Abstract

The application discloses a blood oxygen distribution imaging method and a system based on double-band diffuse reflection space mapping, and relates to the field of non-contact physiological parameter detection for biomedical optical application; and according to the dual-band reflection attenuation ratio, the corresponding effective optical path ratio and the extinction coefficients of oxyhemoglobin and deoxyhemoglobin under the two specific wave bands, utilizing a dual-band inversion model based on the improved lambert-beer law to obtain the preliminary blood oxygen saturation, and generating a two-dimensional blood oxygen saturation distribution map through space regularization optimization. Therefore, the space characterization capability, inversion stability and imaging accuracy of the non-contact blood oxygen imaging can be improved.

Inventors

  • HE YING
  • SUN HUIQIN

Assignees

  • 苏州市立医院

Dates

Publication Date
20260512
Application Date
20260414

Claims (10)

  1. 1. A blood oxygen distribution imaging method based on dual-band diffuse reflection spatial mapping, the method comprising: acquiring a double-band diffuse reflection original image sequence of a tissue region to be detected under two specific wave bands; preprocessing and region segmentation are carried out on the dual-band diffuse reflection original image sequence to obtain a normalized dual-band image and a corresponding region of interest; Estimating the effective optical path ratio of each pixel between the two specific wave bands based on the image gradient information of the normalized dual-band image in the local neighborhood of the corresponding pixel for each pixel in the region of interest to construct a local optical path correction matrix covering the region of interest; For each pixel in the region of interest, calculating preliminary blood oxygen saturation of the pixel by utilizing a preset dual-band inversion model according to reflection attenuation ratio of the normalized dual-band image under the two specific bands and combining the effective optical path ratio of the corresponding pixel in the local optical path correction matrix, wherein the dual-band inversion model is configured to establish a functional mapping relation among the reflection attenuation ratio, the effective optical path ratio, extinction coefficients corresponding to oxyhemoglobin and deoxyhemoglobin under the two specific bands and the preliminary blood oxygen saturation based on an improved lambert-beer law so as to solve and obtain the preliminary blood oxygen saturation of the corresponding pixel; And performing spatial regularization optimization on the preliminary blood oxygen saturation of all pixels in the region of interest to generate a two-dimensional blood oxygen saturation distribution map corresponding to the region of interest.
  2. 2. The method according to claim 1, wherein before the obtaining the sequence of dual-band diffuse reflection original images of the tissue region under test in the two specific bands, the method further comprises determining the center wavelengths corresponding to the two specific bands based on an adaptive wavelength selection mechanism, specifically comprising: Extracting initial skin color characteristics and tissue thickness characteristics corresponding to the tissue region to be detected, and respectively converting the initial skin color characteristics and the tissue thickness characteristics into corresponding melanin index and tissue thickness parameters; calculating sensitivity factors corresponding to each candidate wavelength by combining extinction coefficients of oxyhemoglobin and deoxyhemoglobin in a preset candidate wavelength spectrum library, wherein the sensitivity factors are proportional to absolute values of differences between the extinction coefficients of the oxyhemoglobin and the deoxyhemoglobin under the same candidate wavelength and inversely proportional to standard deviations of system noise commonly influenced by the melanin index and the tissue thickness parameters; Traversing all candidate wavelength combinations in the candidate wavelength spectrum library, and constructing an optimizing function which maximizes the sum of sensitivity factors corresponding to the two candidate wavelengths to be an optimizing target and takes the spectrum distance between the two candidate wavelengths as a constraint, wherein the spectrum distance between the two candidate wavelengths meets a preset constraint condition; And solving the optimizing function to obtain an optimal wavelength combination, and determining two candidate wavelengths corresponding to the optimal wavelength combination as center wavelengths corresponding to the two specific wavebands, wherein the preset constraint condition comprises that the spectrum distance between the two candidate wavelengths is larger than or equal to a minimum spectrum distance threshold value for avoiding the too small local optical path ratio and smaller than or equal to a maximum spectrum distance threshold value for avoiding the too large tissue scattering difference.
  3. 3. The method of claim 1, wherein the preprocessing and region segmentation of the dual-band diffuse reflection original image sequence to obtain a normalized dual-band image and a corresponding region of interest comprises: Acquiring a pre-acquired dark field image and a whiteboard reflection image acquired based on a standard reference whiteboard; Performing flat field correction processing on each image frame in the dual-band diffuse reflection original image sequence by utilizing the dark field image and the white board reflection image to eliminate non-uniform illumination and sensor dark current interference and generate a corresponding correction image sequence, wherein the flat field correction processing comprises the steps of subtracting the dark field intensity of the dark field image from the original reflection intensity of the original image frame, dividing the difference between the white board reflection intensity of the white board reflection image and the dark field intensity to obtain a corresponding normalized reflection intensity; Sub-pixel-level inching compensation based on adjacent frame phase correlation is carried out on the corrected image sequence, and heart rate prior information of the tissue region to be detected is estimated preliminarily from the corrected image sequence after compensation; Extracting a tissue pulsation signal frequency band by utilizing an adaptive band-pass filter for tracking the heart rate prior information so as to filter ambient light noise and low-frequency motion artifacts and generate a normalized dual-band image with time-space alignment; And in the initial contour, performing structural clustering on pixels with homogeneous optical scattering characteristics by using a super-pixel segmentation algorithm to extract a pixel set containing effective physiological tissues as the region of interest, thereby obtaining the normalized dual-band image and the corresponding region of interest.
  4. 4. The method of claim 1, wherein the estimating, for each pixel within the region of interest, an effective optical path ratio of each pixel between the two particular bands based on image gradient information of the normalized dual-band image within a local neighborhood of the corresponding pixel to construct a local optical path correction matrix covering the region of interest comprises: Extracting normalized reflection intensities of each neighborhood pixel in a local neighborhood of the target pixel under the two specific wave bands for each target pixel in the region of interest, and respectively executing local logarithmic transformation to obtain local logarithmic reflection intensity distribution corresponding to the two specific wave bands; for each specific wave band in the two specific wave bands, taking Euclidean distance from each neighborhood pixel in the local neighborhood to the target pixel as an independent variable, taking local logarithmic reflection intensity of the corresponding neighborhood pixel as a dependent variable, establishing a spatial linear regression model, and determining a slope obtained by fitting as a reflection intensity attenuation slope corresponding to the specific wave band, wherein the reflection intensity attenuation slope is used for representing the image gradient information; Based on diffuse reflection diffusion approximation theory, determining an initial optical path ratio estimated value of the target pixel according to the ratio of the reflection intensity attenuation slopes corresponding to the two specific wave bands; Performing iterative regularization smoothing processing on the initial optical path ratio estimated value of each pixel in the region of interest, wherein the iterative regularization smoothing processing comprises the steps of determining Gaussian spatial distance weights based on the spatial distance between pixels in the local neighborhood, determining similarity weights based on image edge features, and performing joint weighted filtering on the initial optical path ratio estimated value by utilizing the Gaussian spatial distance weights and the similarity weights; And determining the iterated and converged optical path ratio value as the effective optical path ratio of the corresponding pixel, and constructing a local optical path correction matrix by the effective optical path ratio of each pixel in the region of interest.
  5. 5. The method of claim 4, wherein the calculating preliminary blood oxygen saturation of the pixel by using a preset dual-band inversion model according to the reflection attenuation ratio of the normalized dual-band image at the two specific bands and the effective optical path ratio of the corresponding pixel in the local optical path correction matrix for each pixel in the region of interest comprises: Extracting normalized reflection intensities of the pixels in the region of interest under a first specific wave band and a second specific wave band in the two specific wave bands, and respectively calculating corresponding tissue absorbance based on negative logarithmic transformation of the normalized reflection intensities; Determining the ratio of the tissue absorbance corresponding to the first specific wave band to the tissue absorbance corresponding to the second specific wave band as the reflection attenuation ratio of the corresponding pixel; substituting the reflection attenuation ratio of the corresponding pixel and the effective optical path ratio corresponding to the pixel in the local optical path correction matrix into the dual-band inversion model, wherein the dual-band inversion model is configured to establish inversion function mapping relations among the reflection attenuation ratio, the effective optical path ratio and extinction coefficients respectively corresponding to oxyhemoglobin and deoxyhemoglobin in the first specific wave band and the second specific wave band and the preliminary blood oxygen saturation based on an improved lambert-beer law, and solve the preliminary blood oxygen saturation of the corresponding pixel by eliminating a total hemoglobin concentration term; The inversion function mapping relation is configured to enable the preliminary blood oxygen saturation to be equal to the ratio of a target molecular term and a target denominator term, wherein the target molecular term is obtained by subtracting the extinction coefficient of deoxyhemoglobin in the first specific wave band from the continuous product of the reflection attenuation ratio, the effective optical path ratio and the extinction coefficient of deoxyhemoglobin in the second specific wave band, the target denominator term is obtained by subtracting the difference of the extinction coefficients of oxyhemoglobin and deoxyhemoglobin in the first specific wave band from the continuous product of the reflection attenuation ratio, the effective optical path ratio and the difference of the extinction coefficients of oxyhemoglobin and deoxyhemoglobin in the second specific wave band, and the effective optical path ratio is the ratio of the effective optical path corresponding to the second specific wave band to the effective optical path corresponding to the first specific wave band.
  6. 6. The method of claim 5, wherein performing spatial regularization optimization on the preliminary blood oxygen saturation of all pixels within the region of interest to generate a two-dimensional blood oxygen saturation profile corresponding to the region of interest comprises: The method comprises the steps of constructing a global variational objective function comprising a data fidelity term and a spatial smoothing regularization term with an edge maintenance weight, wherein the data fidelity term is used for representing a reflection attenuation ratio actual measurement value of a corresponding pixel through error square sum and consistency difference between the reflection attenuation ratio actual measurement value and a theoretical reflection attenuation ratio deduced from the dual-band inversion model according to current blood oxygen saturation and effective optical path ratio, the spatial smoothing regularization term is used for applying weighting penalty to squares of blood oxygen saturation difference values between adjacent pixels under common adjustment of local texture edge maintenance weight and global regularization weight coefficient, and the edge maintenance weight is dynamically determined based on local texture gradient of the normalized dual-band image and is used for reducing excessive smoothing of a tissue structure boundary area; Taking the preliminary blood oxygen saturation corresponding to each pixel as an initial iteration value, and adopting a gradient descent algorithm to carry out minimized iterative updating on the global variation objective function, wherein the blood oxygen saturation of the next iteration round is obtained by updating the blood oxygen saturation of the current iteration round along the negative gradient direction of the global variation objective function at the current blood oxygen saturation and combining with a preset iteration step length; And stopping iteration when the loss reduction amplitude of the global variation objective function is lower than a preset convergence threshold or reaches the maximum iteration number, and determining a blood oxygen saturation result obtained by final convergence as a two-dimensional blood oxygen saturation distribution diagram corresponding to the region of interest.
  7. 7. The method of claim 6, wherein the minimizing the iterative update of the global variational objective function using a gradient descent algorithm with the preliminary blood oxygen saturation corresponding to each pixel as an initial iteration value comprises: Determining the half-width size of a point spread function corresponding to a current optical system, and processing the normalized dual-band image by utilizing a multi-scale vascular enhancement filtering algorithm to determine the local vascular structure scale corresponding to each pixel in the region of interest; Determining a pixel-by-pixel space variable weight regularization coefficient according to a scale proportion relation between the local vascular structure scale and the half-width dimension of the point spread function, and replacing the regularization weight coefficient with the pixel-by-pixel space variable weight regularization coefficient when calculating a space smoothing regularization term and a partial derivative gradient thereof in the global variation objective function, wherein for pixels with the local vascular structure scale smaller than the half-width dimension of the point spread function, the pixel-by-pixel space variable weight regularization coefficient of the corresponding pixels is reduced through a monotonically decreasing mapping function according to the scale proportion; In the iterative updating of the current round, based on the blood oxygen saturation of the current round and the reflection attenuation ratio of the corresponding pixel, synchronously correcting the current effective optical path ratio by combining the dual-band inversion model to obtain the updated effective optical path ratio of the current round; And when the global variation objective function and the partial derivative gradient thereof corresponding to the next round of iteration are calculated, replacing the effective optical path ratio in the global variation objective function with the effective optical path ratio updated by the current round.
  8. 8. A dual-band diffuse reflection spatial mapping-based blood oxygen distribution imaging system, the system comprising: the dual-band image acquisition unit is used for acquiring a dual-band diffuse reflection original image sequence of the tissue region to be detected under two specific bands; the image processing unit is used for preprocessing and region segmentation of the dual-band diffuse reflection original image sequence to obtain a normalized dual-band image and a corresponding region of interest; An optical path correction unit, configured to estimate, for each pixel in the region of interest, an effective optical path ratio of each pixel between the two specific bands based on image gradient information of the normalized dual-band image in a local neighborhood of the corresponding pixel, so as to construct a local optical path correction matrix covering the region of interest; The blood oxygen inversion unit is used for calculating the preliminary blood oxygen saturation of each pixel in the interested region by utilizing a preset dual-band inversion model according to the reflection attenuation ratio of the normalized dual-band image under the two specific bands and the effective optical path ratio of the corresponding pixel in the local optical path correction matrix, wherein the dual-band inversion model is configured to establish the reflection attenuation ratio, the effective optical path ratio and the function mapping relation between the extinction coefficients of oxyhemoglobin and deoxyhemoglobin under the two specific bands and the preliminary blood oxygen saturation respectively based on an improved lambert-beer law so as to solve and obtain the preliminary blood oxygen saturation of the corresponding pixel; And the distribution imaging unit is used for performing spatial regularization optimization on the preliminary blood oxygen saturation of all pixels in the region of interest so as to generate a two-dimensional blood oxygen saturation distribution diagram corresponding to the region of interest.
  9. 9. An electronic device, comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the steps in the method of any one of claims 1-7.
  10. 10. A storage medium having stored thereon a computer program, which when executed by a processor performs the steps of the method according to any of claims 1-7.

Description

Blood oxygen distribution imaging method and system based on double-band diffuse reflection space mapping Technical Field The application relates to the field of non-contact physiological parameter detection for biomedical optical application, in particular to a blood oxygen distribution imaging method and system based on double-band diffuse reflection space mapping. Background The blood oxygen saturation is an important parameter reflecting the oxygen supply state and the physiological function state of human tissues, and has important significance in the scenes of health monitoring, severe nursing, skin lesion evaluation, local tissue blood supply analysis and the like. At present, blood oxygen detection generally mainly depends on a contact type measuring device, blood oxygen information of local parts of a human body is obtained through a light transmission or reflection mode, the mode is mature in application, but is usually required to be in close contact with human skin, local compression and discomfort are easy to bring, and the problem of limited use often exists in special application scenes such as burn nursing, skin abnormality, infection prevention and control and the like. With the development of non-contact physiological parameter detection technology, remote blood oxygen detection on a target area by utilizing an optical imaging mode is gradually attracting attention. The blood oxygen state is deduced by collecting the reflected light signals of the surface of the human body, so that the blood oxygen state has certain advantages in the aspects of reducing contact risks and improving use convenience. However, from the current technical development, the non-contact blood oxygen detection still faces obvious limitations in practical application, namely, one type of scheme depends on complex imaging equipment such as multispectral or hyperspectral and the like, but can acquire rich spectral information, but the problems of complex equipment, higher cost, lower data acquisition efficiency, complex correction process and the like generally exist, the other type of scheme is simplified based on common imaging equipment, and the hardware threshold is relatively lower, but the requirements of high-precision application are generally difficult to meet in terms of signal quality, anti-interference capability and measurement stability. In addition, the current non-contact blood oxygen detection technology has limited capacity for representing the spatial difference of the blood oxygen state in the target area, and has room for further improvement in measurement stability and imaging accuracy under the influence of complex tissue optical characteristics. Disclosure of Invention The application provides a blood oxygen distribution imaging method, a system, a storage medium, a computer program product and electronic equipment based on dual-band diffuse reflection space mapping, which are used for at least solving the problems of insufficient measurement accuracy and space distribution characterization capability in non-contact blood oxygen detection in the prior art. In a first aspect, an embodiment of the present application provides a blood oxygen distribution imaging method based on dual-band diffuse reflection spatial mapping, where the method includes obtaining a dual-band diffuse reflection original image sequence of a tissue region to be measured under two specific bands; the method comprises the steps of preprocessing and region segmentation of a dual-band diffuse reflection original image sequence to obtain a normalized dual-band image and a corresponding region of interest, estimating the effective optical path ratio of each pixel between two specific bands based on image gradient information of the normalized dual-band image in the local neighborhood of the corresponding pixel for each pixel in the region of interest to construct a local optical path correction matrix covering the region of interest, solving a function relation between extinction coefficients and the preliminary blood oxygen saturation of each pixel in the region of interest according to the reflection attenuation ratio of the normalized dual-band image in the two specific bands, combining the effective optical path ratio of the corresponding pixel in the local optical path correction matrix, calculating the preliminary blood oxygen saturation of the pixel by using a preset dual-band inversion model, and the dual-band inversion model is configured to establish a function relation among the reflection attenuation ratio, the effective optical path ratio, oxyhemoglobin and deoxyhemoglobin under the two specific bands respectively, so as to obtain a preliminary blood oxygen saturation map of the corresponding pixel, and performing preliminary blood oxygen saturation of the corresponding region of interest in a two-dimensional saturated region of interest, and performing preliminary saturation of all the pixels in the preliminary sat