Double-layer medium wide spectrum parameter simultaneous reconstruction method based on spectrum optimization
By using spectral optimization methods and algorithms, combined with light field cameras and prior information on spectral dimensions, the problem of low reconstruction accuracy of multilayer media under traditional contact measurement methods was solved, and efficient and accurate reconstruction of optical parameters of dual-layer media was achieved.
Patent Information
- Application Number
- CN202511358596.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-23
- Publication Date
- 2026-01-23
AI Technical Summary
In existing technologies, when reconstructing wide optical parameters based on multilayer media models, the detection radiation light signal is limited by traditional contact measurement methods, resulting in low reconstruction accuracy.
A method for simultaneous reconstruction of broadband spectral parameters of a two-layer medium based on spectral optimization is adopted. This method combines a light field camera to accurately detect the reflected radiation signal at the medium boundary, and utilizes the steepest descent method and conjugate gradient method to optimize the algorithm. By combining prior information on the spectral dimension, the method achieves efficient reconstruction of the optical parameters inside the medium.
It significantly improves the reconstruction accuracy and efficiency of bilayer media, especially the reconstruction effect of the absorption coefficient and reduced scattering coefficient of the bottom medium, realizing non-destructive detection inside the medium.
Smart Images
Figure CN121389422A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application relates to the technical field of spectral imaging, in particular to a wide-spectrum parameter simultaneous reconstruction method for a double-layer medium based on spectral optimization. BACKGROUND
[0002] The research on the radiation inverse problem in a participating medium has become one of key problems in the fields of atmospheric pollution detection, biomedical imaging, industrial quality control and the like. The participating medium widely exists in various industrial fields. When laser is incident to the participating medium, the photon can be regarded as an information medium transmitted in the medium, and the absorption and scattering of the photon by the medium are researched. The radiation transmission of the photon is determined by the unique optical characteristic distribution in the medium, and is crucial for the diagnosis and detection of the interior of the participating medium.
[0003] Generally, the optical parameters in the medium interior are difficult to be directly measured, the effective information in the radiation light signal of the medium boundary can be detected, the radiation light signal is analyzed and processed in combination with an inverse problem algorithm, the optical parameters in the interior of the participating medium are indirectly reconstructed, and the nondestructive detection of the medium interior is realized. However, the research on the optical parameter reconstruction of the participating medium still faces many problems at present, especially when the wide optical parameter reconstruction is carried out based on a multi-layer medium model, the detected radiation light signal is limited by the traditional contact type measurement mode, and the reconstruction accuracy is low. SUMMARY
[0004] The application aims at the problem that the detected radiation light signal is limited by the traditional contact type measurement mode and the reconstruction accuracy is low when the wide optical parameter reconstruction is carried out based on the multi-layer medium model, and provides a wide-spectrum parameter simultaneous reconstruction method for a double-layer medium based on spectral optimization.
[0005] The technical scheme adopted by the application to solve the above technical problem is:
[0006] The wide-spectrum parameter simultaneous reconstruction method for a double-layer medium based on spectral optimization comprises the following steps:
[0007] Step 1: a short pulse continuous light source 1 is started, the laser emitted by the short pulse continuous light source 1 passes through a dichroic mirror 2 and an attenuator 3 in sequence, single-color pulse laser of different wavelengths is obtained, all the single-color pulse laser is screened through a filter sheet screen 4, the pulse laser of a specific wavelength is selected, then the selected pulse laser passes through a fiber coupler 5 and an optical fiber fixed on the surface of a participating medium 7, is vertically incident to the participating medium 7, finally, a reflection radiation measurement signal of different detection positions on the surface of the participating medium 7 is collected by using a light field camera 6, and the participating medium 7 is a double-layer medium;
[0008] Step 2: Obtain the initial value of the spectral parameter of the participating medium 7, and bring the initial value of the spectral parameter into the time-domain diffusion approximation equation to solve, to obtain all the reflected radiation at this wavelength. After all the reflected radiation is taken as a boundary signal, combine all the reflected radiation measurement signals in step 1 to construct an initial objective function;
[0009] Step 3: Based on the initial objective function, and according to the steepest descent method, update the spectral parameter distribution value in the participating medium 7. If the iteration gradient in the iteration process is less than the set minimum value , then end the iteration, and obtain the spectral parameter distribution value ;
[0010] Step 4: The short pulse continuous light source 1 emits times of laser, and repeats steps 1 to 3 to obtain spectral parameter distribution values at different wavelengths;
[0011] Step 5: Obtain the upper medium absorption coefficient , the upper medium scattering coefficient , the lower medium absorption coefficient , and the lower medium scattering coefficient corresponding to the spectral parameter distribution value at each wavelength, and use the upper medium absorption coefficient , the upper medium scattering coefficient , the lower medium absorption coefficient , and the lower medium scattering coefficient corresponding to the spectral parameter distribution value at all wavelengths to construct the spectral parameter under a wide spectrum, which is represented as:
[0012] ,
[0013] ,
[0014] Among them, =1,2,..., , represents the spectral parameter under a wide spectrum;
[0015] Step 6: Substitute the spectral parameter under a wide spectrum into the time-domain diffusion approximation equation and solve to obtain new boundary signals , then combine the new boundary signals with all the reflected radiation measurement signals in step 1 to establish an objective function ;
[0016] Step 7: Based on the objective function , introduce spectral prior information, and according to the conjugate gradient method, update the spectral parameter distribution value in the participating medium 7. If the objective function value is less than the minimum value , then the iteration is ended, and the spectral parameter distribution value is obtained ;
[0017] Step 8: the spectral parameter distribution value is brought into the time-domain diffusion approximation equation of the double-layer medium to obtain the reflected radiation signal of the medium boundary , then the objective function is updated by using the reflected radiation signal , and the objective function is obtained, if the objective function value in the objective function is less than the minimum value , the iteration is ended, and Step 9 is executed, otherwise, the spectral parameter distribution is replaced with the spectral parameter under the wide spectrum in Step 6, and the iteration is performed;
[0018] Step 9: the spectral parameter distribution value obtained at present is taken as the wide spectrum parameter reconstruction result.
[0019] Further, the spectral parameter initial value is represented as:
[0020] ,
[0021] wherein, μa (λ) and μ's (λ) represent the wide spectrum absorption spectrum and the reduced scattering spectrum in the double-layer medium model respectively. ,
[0022] Further, the spectral parameter distribution value is represented as:
[0023] ,
[0024] wherein, Δt represents the iteration step length, g represents the iteration gradient, and represents the spectral parameter distribution value of the i-th iteration.
[0025] Further, the time-domain diffusion approximation equation is represented as:
[0026] ,
[0027] wherein, r represents the position, c represents the light speed in the tissue medium, t represents the time, D represents the diffusion coefficient, μa represents the absorption coefficient, F represents the projected radiation force, and n represents the refractive index. represents a radiation source term introduced by incident laser, represents a Hamiltonian operator.
[0028] Further, the diffusion coefficient is represented as:
[0029] ,
[0030] wherein, represents an effective scattering coefficient.
[0031] Further, the effective scattering coefficient is represented as:
[0032] ,
[0033] wherein, represents a scattering asymmetry factor, represents a scattering coefficient.
[0034] Further, the initial objective function is represented as:
[0035] ,
[0036] wherein, represents an initial objective function, represents a spectral parameter to be measured, represents an i-th pulse laser band, represents an i-th detection position corresponding to the i-th pulse laser band, represents a boundary signal calculated according to a spectral parameter and a time-domain diffusion approximation equation under an i-th band. Further, the spectral parameter distribution value is represented as:
[0037] ,
[0038] wherein, represents a step size of an i-th iteration of the entire spectral parameter,
[0039] represents a descent direction corresponding to the i-th iteration, represents a spectral parameter distribution value obtained by an i-1th iteration. Further, the objective function is represented as:
[0040] Further, the objective function is represented as:
[0041] ,
[0042] wherein, represents the boundary signal calculated according to the spectral parameter and the time-domain diffusion approximation equation, represents the prior information introduced by the spectral dimension, i.e. the spectral prior information.
[0043] Further, the prior information introduced by the spectral dimension is represented as:
[0044] ,
[0045] wherein, and represents the power-law coefficient, represents the laser wavelength band, represents the component content coefficient, represents the medium component spectral characteristic parameter, , respectively represent the wide-spectrum absorption spectrum and the reduced scattering spectrum in the double-layer medium model.
[0046] The beneficial effects of the present application are:
[0047] On the basis of the traditional participatory medium optical parameter measurement and reconstruction method, the present application introduces the more accurate detection of the reflected radiometric signal of the medium boundary by the light field camera; and combines the reconstruction optimization algorithm of "steepest descent method to find initial point + conjugate gradient method to find optimal solution", which greatly accelerates the efficiency of the accurate reconstruction of the double-layer medium, breaks the limitation of the traditional contact measurement method, and improves the reconstruction accuracy; considering the extremely low sensitivity of the radiometric signal to the reduced scattering coefficient of the bottom layer medium, the present application introduces the prior information of the spectral dimension, which improves the reconstruction results of the absorption coefficient and the reduced scattering coefficient of the double-layer medium, especially the bottom layer medium, and realizes the simultaneous reconstruction of the absorption spectrum and the reduced scattering spectrum of the double-layer medium. BRIEF DESCRIPTION OF DRAWINGS
[0048] Figure 1 is a schematic diagram of the influence of the change of the optical parameters of the bottom layer of the double-layer medium on the reflected signal Figure 1 ;
[0049] Figure 2 is a schematic diagram of the influence of the change of the optical parameters of the bottom layer of the double-layer medium on the reflected signal Figure 2 ;
[0050] Figure 3 is a schematic diagram of the reconstruction method structure of the present application;
[0051] Figure 4 is a flowchart of the reconstruction method of the present application. DETAILED DESCRIPTION
[0052] It should be particularly noted that the various embodiments disclosed in the present application can be combined with each other in the case of no conflict.
[0053] Specific implementation one: the double-layer medium wide spectrum parameter simultaneous reconstruction method based on spectrum optimization in the embodiment, the method comprises the following steps:
[0054] Step one: turn on the short pulse continuous light source 1, decompose the laser into monochromatic light of different wavelengths through the dichroic mirror 2 and the attenuator 3, and all the monochromatic light passes through the filter screen 4 composed of a special filter, and the pulse laser under the specific wavelength is selected. Then, the laser passes through the optical fiber coupler 5 and the optical fiber vertically incident to the participating medium 7 fixed on the medium surface, and the light field camera 6 collects the reflection radiation measurement signals of the participating medium 7 at different positions corresponding to the wavelength;
[0055] Step two: repeat step one, the short pulse continuous light source 1 emits a total of times of laser, and through the screening of the filter screen 4, a total of pulse lasers of specific spectrum wide spectrum can be obtained, and the measurement signals of the light field camera 6 collected at bands corresponding to positions are sent to the data acquisition and processing system 8 for processing, so as to obtain the reflection radiation measurement signals of the participating medium 7 at different positions , , , , , , , ,
[0056] Step three: assuming that the participating medium is a uniform medium, setting the initial value of the optical parameter as , , , , , ,
[0057] Step four: according to the steepest descent method, the distribution value of the optical parameter in the participating medium is updated, and the distribution value of the optical parameter of the first +1 step iteration is obtained , , , , This represents the iterative gradient. If the iterative gradient is less than the set minimum value during the iteration process... If so, the iteration ends;
[0058] Step 5: Repeat steps 3 and 4 to obtain... Reconstructed values of all optical parameters under different wavebands , =1,2,… ,Will , These are used as the initial values for the prediction of the broadband absorption spectrum and the reduced scattering spectrum in the two-layer medium model, respectively.
[0059] Step Six: Substitute the predicted broadband spectral parameters into the time-domain diffusion approximation equation and solve to obtain the new boundary signal. Then, it is compared with the measurement signal obtained in step one. Establish the objective function together Furthermore, prior information based on the spectral dimension is introduced into the newly established objective function;
[0060] Step 7: Construct a reconstruction model based on the conjugate gradient method combined with prior spectral information, and update the distribution values of spectral parameters in the participating medium. If the objective function value is less than the minimum value... Then the iteration ends, and the result is the first iteration. Distribution of spectral parameters in step iteration And calculate the corresponding spectral prior information coefficients. , and ;
[0061] Step 8: Based on the spectral parameter distribution obtained in Step 7 By substituting the approximate time-domain diffusion equation of the two-layer medium into the calculation, the reflected radiation signal at the medium boundary is obtained. Calculate the objective function If the objective function The objective function value is less than the minimum value. Proceed to step nine; otherwise, proceed to step six.
[0062] Step 9: Input the spectral parameters obtained in the current iteration The inversion process ends with the reconstruction results of the broadband parameters.
[0063] Participating medium spectral parameter values Here, x includes the absorption coefficients of the upper medium corresponding to all broadband spectra. upper medium scattering coefficient absorption coefficient of the lower layer medium and the scattering coefficient of the underlying medium .Right now: , .
[0064] The time-domain diffusion approximation equation is solved by an analytical solution method, wherein the expression of the time-domain diffusion approximation equation is as follows:
[0065] ,
[0066] wherein, denotes a position, denotes a light speed in a tissue medium, denotes time, denotes a diffusion coefficient, denotes an absorption coefficient, denotes a projected radiation force, denotes a refractive index, denotes a radiation source term introduced by incident laser, denotes a Hamiltonian operator.
[0067] The diffusion coefficient is expressed as:
[0068] ,
[0069] wherein, denotes an effective scattering coefficient.
[0070] The effective scattering coefficient is expressed as:
[0071] ,
[0072] wherein, denotes a scattering asymmetry factor, denotes a scattering coefficient.
[0073] The initial objective function is:
[0074] ,
[0075] wherein, is an optical parameter to be measured, denotes a measurement signal, denotes a boundary signal calculated according to the optical parameter and the time-domain diffusion approximation equation at the i-th wave band.
[0076] The distribution value of the spectral parameter in the participating medium 8 is updated by using the conjugate gradient method in step seven, and is expressed as:
[0077] ,
[0078] wherein, Step length of the kth iteration of the whole spectrum parameter, Descent direction corresponding to the kth iteration.
[0079] Objective function described in step eight :
[0080] ,
[0081] wherein, Boundary signal calculated according to the spectrum parameter x and the time domain diffusion approximation equation, Prior information introduced in the spectral dimension.
[0082] Specific description of the prior information in the spectral dimension in step six:
[0083] According to the underlying scattering coefficient and the absorption coefficient of the wide optical parameter of the participating medium, the wavelength variation conforms to the following scale invariance relationship and power law relationship: , Therefore, a penalty term is added in the objective function:
[0084] ,
[0085] wherein, and Both are power law coefficients, and their values depend on the size and number density of the scattering center; Corresponding laser band; Component content coefficient, the value of which is related to the distribution of chromophores inside the medium. Medium component spectral characteristic parameter.
[0086] The application uses a light field camera instead of a traditional contact photoelectric detector, and through the setting of optical elements such as a microlens array in front of the sensor, the intensity, direction and position information of the radiation are recorded at the same time, which significantly improves the detection accuracy of the boundary radiation signal; at the same time, in the reconstruction of the inverse problem, aiming at the problem that the reduced scattering coefficient of the bottom layer medium in the double-layer medium model is less sensitive and difficult to reconstruct, the application finds the initial point of the feasible region based on the steepest descent method, combines the conjugate gradient method to find the optimal solution, improves the reconstruction efficiency of complex multi-objective optimization problems, further constructs the constraint condition based on the scale invariance and spectral continuity of the participating medium, improves the reconstruction result of the absorption coefficient and the reduced scattering coefficient of the double-layer medium, especially the bottom layer medium, improves the accuracy and robustness of the optical parameter reconstruction, realizes the simultaneous reconstruction of the absorption spectrum and the reduced scattering spectrum of the double-layer medium, and completes the accurate and non-destructive detection inside the participating medium.
[0087] The application can be applied to the fields of wide-spectrum nondestructive detection, combustion process diagnosis, infrared remote sensing detection, biomedical imaging and the like to provide key technical support for accurate reconstruction of internal optical parameters of a participating medium.
[0088] It should be noted that the specific embodiments are only an explanation and illustration of the technical solutions of the present application, and cannot limit the protection scope. Any partial change made according to the claims and the specification of the present application shall still fall within the protection scope of the present application.
Claims
1. A method for simultaneous reconstruction of wide spectrum parameters of a two-layer medium based on spectral optimization, characterized in that The method comprises the following steps: Step 1: turn on the short-pulse continuous light source (1), the laser emitted by the short-pulse continuous light source (1) passes through the dichroic mirror (2) and the attenuator (3) in turn to obtain monochromatic pulse lasers of different wavelengths, all the monochromatic pulse lasers are screened through the filter screen (4) to select pulse lasers of a specific wavelength, then the selected pulse lasers are coupled through the optical fiber coupler (5) and the optical fiber fixed on the surface of the participating medium (7), and are vertically incident on the participating medium (7), finally, the reflected radiation measurement signals of different detection positions on the surface of the participating medium (7) are collected by using the light field camera (6), and the participating medium (7) is a double-layer medium; Step 2: obtain the initial value of the spectral parameter of the participating medium (7), and input the initial value of the spectral parameter into the time-domain diffusion approximation equation to solve, so as to obtain all the reflected radiations at the wavelength, after all the reflected radiations are taken as boundary signals, the initial objective function is constructed in combination with all the reflected radiation measurement signals in step 1; Step 3: updating the spectral parameter distribution value in the participating medium (7) based on the initial objective function and according to the steepest descent method, if the iteration gradient in the iteration process is less than a set minimum value , then ending the iteration to obtain the spectral parameter distribution value ; Step 4: the short-pulse continuous light source (1) emits sub-laser, repeating steps 1 to 3 to obtain spectrum parameter distribution values at different wavelengths; Step 5: obtaining the upper layer medium absorption coefficient corresponding to the spectral parameter distribution value at each wavelength , the upper layer medium scattering coefficient , the lower layer medium absorption coefficient , and the lower layer medium scattering coefficient , and using the upper layer medium absorption coefficient corresponding to the spectral parameter distribution value at all wavelengths , the upper layer medium scattering coefficient , the lower layer medium absorption coefficient , and the lower layer medium scattering coefficient to construct the spectral parameter under a wide spectrum, and the spectral parameter under the wide spectrum is represented as: ， ， wherein = 1, 2,... , denotes a spectral parameter under a broad spectrum; Step 6: Substitute the spectral parameters under the wide spectrum into the time-domain diffusion approximation equation and solve to obtain a new boundary signal Then combine the new boundary signal with all the reflected radiation measurement signals in Step 1 to establish an objective function ; Step 7: based on the objective function , the spectral prior information is introduced, and the spectral parameter distribution value in the participating medium (7) is updated according to the conjugate gradient method. If the objective function value is less than the minimum value , the iteration is ended, and the spectral parameter distribution value is obtained . Step 8: input the spectral parameter distribution value into the two-layer medium time-domain diffusion approximation equation to obtain the reflected radiation signal of the medium boundary , then update the objective function using the reflected radiation signal , obtain the objective function , if the objective function value in the objective function is less than the minimum value , end the iteration, and execute Step 9, otherwise, replace the spectral parameter under the wide spectrum in Step 6 with the spectral parameter distribution , and iterate; Step 9: take the obtained spectral parameter distribution value as the wide-spectrum parameter reconstruction result.
2. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 1, characterized in that The initial value of the spectral parameter is represented as: ， wherein , respectively represent the wide-spectrum absorption spectrum and the reduced scattering spectrum in the two-layer medium model.
3. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 1, wherein The spectral parameter distribution values is represented as: ， wherein, denotes the iteration step, denotes the iteration gradient, denotes the distribution value of the spectral parameter of the step iteration.
4. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 3, characterized in that The time-domain diffusion approximation equation is represented as: ， wherein, represents the position, represents the speed of light in the tissue medium, represents the time, represents the diffusion coefficient, represents the absorption coefficient, represents the projected radiation force, represents the refractive index, represents the radiation source term introduced by the incident laser, represents the Hamiltonian operator.
5. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 4, characterized in that The diffusion coefficient is expressed as: ， wherein denotes the effective scattering coefficient.
6. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 5, characterized in that The effective scattering coefficient is expressed as: ， wherein denotes the scattering asymmetry factor, denotes the scattering coefficient.
7. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 6, characterized in that The initial objective function is represented as: ， in, Describes the initial objective function. Indicates the spectral parameters to be measured. Indicates the first The corresponding pulse laser band One detection location, Indicates the first Based on spectral parameters in each band The boundary signal is calculated using the time-domain diffusion approximation equation.
8. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 7, wherein The spectral parameter distribution values is represented as: ， wherein, represents the step size of the entire spectrum parameter for the th iteration, represents the descent direction corresponding to the th iteration, represents the spectrum parameter distribution value obtained from the -1th iteration.
9. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 8, wherein The objective function is represented as: ， wherein denotes a boundary signal calculated from a spectral parameter and a time-domain diffusion approximation equation, denotes a prior information introduced by the spectral dimension, i.e. spectral prior information.
10. The dual-layer medium based spectral optimization method for simultaneous reconstruction of wide spectral parameters according to claim 9, wherein the prior information introduced by the spectral dimension is represented as: ， wherein, and represents a power law coefficient, represents a laser wavelength, represents a component content coefficient, represents a medium component spectral characteristic parameter, , respectively represent a wide-spectrum absorption spectrum and a reduced scattering spectrum in a two-layer medium model.