A bandwidth correction method and system for LED multispectral imaging system

Abstract

The present invention provides a method for correcting the bandwidth of an LED multispectral imaging system, comprising obtaining a standard reflectance of a predicted color card under the LED multispectral imaging system and its reconstructed spectral reflectance, and setting the spectral reflectance reconstructed from the predicted color card as the spectral reflectance to be corrected; defining a bandwidth function based on the bandwidth of all LED light sources in the LED multispectral imaging system; performing iterative bandwidth correction on the spectral reflectance to be corrected using the Richardson-Lucy iterative method according to the bandwidth function until a predetermined iteration termination condition is met; after the iterative calculation is completed, outputting the final correction bandwidth and step size as the optimal bandwidth correction parameters of the LED multispectral imaging system. The present invention also provides a bandwidth correction system for an LED multispectral imaging system. The implementation of the present invention aims to solve the problem that the accuracy of spectral reflectance reconstruction in an LED multispectral imaging system is difficult to further improve.

Description

Technical Field

[0001] The present invention relates to the technical field of spectral imaging, and in particular to a bandwidth correction method and system for an LED multi-spectral imaging system. Background Art

[0002] With the rapid development of science and technology and people's constant pursuit of a better quality of life, color measurement and reproduction technologies are playing an increasingly important role. Traditional spectrophotometers can measure spectral reflectance, but their inherent structural characteristics limit their application in many scenarios, such as the inability to measure powdered, irregular, or tiny objects. Thanks to their non-contact design, multispectral imaging systems can meet measurement requirements without being restricted by these measurement environments.

[0003] Furthermore, while modern spectrometer technology has made significant progress, measured spectra are still subject to errors introduced by instrument bandwidth, stray light, and noise. Ideally, the LED light source in a multispectral imaging system should have an extremely narrow bandwidth, similar to a pulse signal. However, in practice, LEDs used have varying linewidths, affecting the resulting spectrum and causing the reconstructed spectral reflectance to deviate from the actual value. Therefore, bandwidth correction is also necessary.

[0004] When color measurement devices based on LED multispectral imaging systems struggle to effectively improve the spectral reflectance reconstruction accuracy through hardware modifications, bandwidth correction can further improve this accuracy without changing the hardware structure. However, existing technologies are mostly applied to bandwidth correction of spectrometers, and no research has yet applied this to bandwidth correction of LED multispectral imaging systems.

[0005] Therefore, it is necessary to propose a bandwidth correction method for LED multispectral imaging system, aiming to solve the problem that the spectral reflectance reconstruction accuracy of LED multispectral imaging system is difficult to further improve, which has important practical significance for improving the bandwidth correction accuracy of LED multispectral imaging system. Summary of the Invention

[0006] The technical problem to be solved by the embodiments of the present invention is to provide a method and system for bandwidth correction of an LED multi-spectral imaging system, aiming to solve the problem that the accuracy of spectral reflectance reconstruction of the LED multi-spectral imaging system is difficult to further improve.

[0007] In order to solve the above technical problems, an embodiment of the present invention provides a method for bandwidth correction of an LED multispectral imaging system, the method comprising the following steps:

[0008] Obtaining the standard reflectance and spectral reflectance of the predicted color card under the LED multispectral imaging system, and setting the spectral reflectance of the predicted color card as the spectral reflectance to be corrected;

[0009] Based on the bandwidth of all LED light sources in the LED multispectral imaging system, a bandwidth function is defined;

[0010] According to the bandwidth function, using the Richardson-Lucy iterative method, performing bandwidth iterative correction on the spectral reflectance to be corrected until a predetermined iteration end condition is met;

[0011] After the iterative calculation is completed, the final correction bandwidth and step size are output as the optimal bandwidth correction parameters of the LED multi-spectral imaging system.

[0012] Wherein, the LED light source in the LED multi-spectral imaging system comes from the LED active lighting system; wherein,

[0013] The LED active lighting system is installed on top of the pre-installed lighting cabinet and above the camera. It consists of 9 identical circuit boards, and each circuit board has M LED monochromatic light sources evenly arranged; the peak wavelengths of the M LEDs are evenly distributed in the range of 350nm to 780nm;

[0014] The camera is installed in the middle of the top of the lighting cabinet and below the LED active lighting system. A light-diffusing plate is installed on the lens plane, and the lens is aimed at a known training color card or a predicted color card placed horizontally at the bottom of the lighting cabinet.

[0015] The bandwidth function is constructed based on the full width at half maximum obtained from the relative spectral power distribution of each LED light source;

[0016] The bandwidth function is a Gaussian function or a symmetrical triangular function; if the bandwidth function is a Gaussian function, the variance of each Gaussian function is obtained according to the full width at half maximum; if the bandwidth function is a symmetrical triangular function, the upper and lower limits of the horizontal coordinate of the piecewise function are obtained according to the full width at half maximum.

[0017] The step of performing bandwidth iterative correction on the spectral reflectance to be corrected using the Richardson-Lucy iterative method according to the bandwidth function until a predetermined iteration end condition is met specifically includes:

[0018] (1.1) Initialization parameter m = m0. m0 can be 0.01;

[0019] (1.2) Initialization parameter n = n0, n0 can be 0.01;

[0020] (1.3) The line width of the bandwidth function under the same LED light source channel involved in the calculation is set to m times the FWHM of the initial test function, and the correction step size is set to n times the corresponding line width. The correction step size is δ = FWHM*m*n;

[0021] (1.4) Calculate the number of unilateral correction points N1 = [1 / n], N2 = [1 / n]. Where [·] represents the largest integer not greater than “·”, N1 represents the number of correction points to the left of the correction point, and N2 represents the number of correction points to the right of the correction point. Where N1 = N2, the total number of points in the bandwidth function is N = N1 + N2 + 1.

[0022] (1.5) The discretized bandwidth function is expressed as T stands for transpose;

[0023] (1.6) Calculate the first estimate of the measured spectrum:

[0024]

[0025] in, is the initial estimated spectrum, which is the reflectance of the spectrum to be corrected; Mirror image of bandwidth function; is the spectrum estimate at i sampling intervals (δ) to the right of the central wavelength point; * represents convolution; K is the number of points to be corrected, which is the number of LED types;

[0026] (1.7) Calculate the correction factor:

[0027]

[0028] Among them, M(λ k ) is the measured spectrum, i.e. the obtained spectral reflectance to be corrected;

[0029] (1.8) Convolve the correction factor with the bandwidth function to obtain the correction term:

[0030]

[0031] (1.9) The correction term is combined with the previous estimate Multiply them together to get a new estimate:

[0032]

[0033] The overall calculation process can also be expressed as follows:

[0034]

[0035] in is the updated value, is the value before update, r=0,1,2…, r+1 is the number of iterations;

[0036] (1.10) Return to step (1.6) until the set iteration end point is reached; the end point can be set to the difference between the two iterations before and after or the number of iterations, which can be set to or r+1≥100;

[0037] (1.11) Update the parameter n = n + e and return to step (1.3) until n increases to 1; where e can be 0.01;

[0038] (1.12) Update the parameter m = m + e and return to step (1.2) until m increases to 2; where e can be 0.01;

[0039] Wherein, the method further comprises:

[0040] In the Richardson-Lucy iterative method, the values of the initialization parameters m and n are adjusted multiple times;

[0041] The bandwidth correction result obtained when the spectral reflectance to be corrected is subjected to bandwidth iterative correction using the Richardson-Lucy iterative method with each parameter adjustment is compared with the standard reflectance of the predicted color card to obtain the optimal values of m and n.

[0042] The step of comparing the bandwidth correction result obtained by performing bandwidth iterative correction on the spectral reflectance to be corrected using the Richardson-Lucy iterative method with each parameter adjustment with the standard reflectance of the predicted color card to obtain the optimal values of m and n specifically includes:

[0043] Interpolating the corrected reflectance under each adjusted value of m and n to the same wavelength range of the spectrophotometer; wherein the interpolation algorithm uses Lagrange fifth-order interpolation or cubic spline interpolation;

[0044] The interpolated corrected reflectivity is compared with the standard reflectivity of the predicted color block in the predicted color card to calculate the root mean square error value; or, the interpolated corrected reflectivity and the standard reflectivity of the predicted color block are respectively converted into color values under a standard light source, and the color difference value between the corrected color and the standard color is calculated using a color difference formula; wherein the color difference formula can be a CIE76 color difference formula or a CIEDE2000 color difference formula.

[0045] By comparing the root mean square error or color difference value under different values of m and n, the optimal m and n values are obtained as the optimal bandwidth correction parameter combination of the LED multi-spectral imaging system.

[0046] The spectral reflectance of the predicted color card is obtained by performing the following steps:

[0047] Selecting a training color card and a prediction color card; wherein the training color card and the prediction color card are both composed of color blocks containing multiple hues and chromas;

[0048] Using the LED multispectral imaging system to shoot, obtain the camera response of each training color block in the training color chart and the camera response of each predicted color block in the predicted color chart;

[0049] Obtaining a standard reflectance of each training color block and each predicted color block; wherein the standard reflectance can be obtained by an instrument such as a spectrophotometer;

[0050] Use the interpolation algorithm to interpolate the standard reflectance of each training color block and each predicted color block to the reflectance under the corresponding LED channel;

[0051] Establishing a conversion relationship between the camera response of the training color block and the standard reflectivity; wherein the conversion relationship is constructed using one or more methods of Wiener estimation method, pseudo-inverse method, and neural network;

[0052] The spectral reflectance of the predicted color card is calculated based on the conversion relationship.

[0053] An embodiment of the present invention further provides a bandwidth correction system for an LED multi-spectral imaging system, comprising:

[0054] The spectral reflectance acquisition unit to be corrected is used to obtain the standard reflectance and spectral reflectance of the predicted color card under the LED multi-spectral imaging system, and set the spectral reflectance of the predicted color card as the spectral reflectance to be corrected;

[0055] A bandwidth function definition unit, configured to define a bandwidth function based on the bandwidths of all LED light sources in the LED multi-spectral imaging system;

[0056] a bandwidth correction iteration unit, configured to perform bandwidth iterative correction on the spectral reflectance to be corrected according to the bandwidth function using a Richardson-Lucy iteration method until a predetermined iteration end condition is met;

[0057] The correction bandwidth output unit is used to output the final correction bandwidth and step size as the optimal bandwidth correction parameters of the LED multi-spectral imaging system after the iterative calculation is completed.

[0058] Wherein, the LED light source in the LED multi-spectral imaging system comes from the LED active lighting system; wherein,

[0059] The LED active lighting system is installed on top of the pre-installed lighting cabinet and above the camera. It consists of 9 identical circuit boards, and each circuit board has M LED monochromatic light sources evenly arranged; the peak wavelengths of the M LEDs are evenly distributed in the range of 350nm to 780nm;

[0060] The camera is installed in the middle of the top of the lighting cabinet and below the LED active lighting system. A light-diffusing plate is installed on the lens plane, and the lens is aimed at a known training color card or a predicted color card placed horizontally at the bottom of the lighting cabinet.

[0061] The bandwidth function is constructed based on the full width at half maximum obtained from the relative spectral power distribution of each LED light source;

[0062] The bandwidth function is a Gaussian function or a symmetrical triangular function; if the bandwidth function is a Gaussian function, the variance of each Gaussian function is obtained according to the full width at half maximum; if the bandwidth function is a symmetrical triangular function, the upper and lower limits of the horizontal coordinate of the piecewise function are obtained according to the full width at half maximum.

[0063] The implementation of the embodiments of the present invention has the following beneficial effects:

[0064] The present invention defines a bandwidth function based on the bandwidth of all LED light sources in the LED multispectral imaging system. The Richardson-Lucy iterative method is used to perform bandwidth iterative correction on the spectral reflectance to be corrected, and the optimal correction parameters of the LED multispectral imaging system are output, thereby solving the problem that the accuracy of spectral reflectance reconstruction of the LED multispectral imaging system is difficult to further improve. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying any creative work, other drawings obtained based on these drawings still fall within the scope of the present invention.

[0066] Figure 1 A flowchart of a bandwidth correction method for an LED multi-spectral imaging system provided by an embodiment of the present invention;

[0067] Figure 2 A schematic diagram of the installation of an LED active lighting system and a camera in a lighting cabinet in a bandwidth correction method for an LED multispectral imaging system provided by an embodiment of the present invention;

[0068] Figure 3 A diagram showing the relative spectral power distribution of LEDs in a bandwidth correction method for an LED multi-spectral imaging system provided by an embodiment of the present invention;

[0069] Figure 4 A distribution diagram of a Gaussian function in a bandwidth correction method for an LED multi-spectral imaging system provided by an embodiment of the present invention;

[0070] Figure 5 A diagram showing the effect of different m and n values on the average color difference under the CIE76 color difference formula under the Wiener estimation method in a bandwidth correction method for an LED multi-spectral imaging system provided by an embodiment of the present invention;

[0071] Figure 6 A schematic structural diagram of a bandwidth correction system for an LED multi-spectral imaging system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be described in further detail below with reference to the accompanying drawings.

[0073] like Figure 1 FIG. 1 is a method for correcting bandwidth of an LED multi-spectral imaging system according to an embodiment of the present invention, the method comprising the following steps:

[0074] Step S1, obtaining the standard reflectivity and spectral reflectivity of the predicted color card under the LED multi-spectral imaging system, and setting the spectral reflectivity of the predicted color card as the spectral reflectivity to be corrected;

[0075] Step S2: defining a bandwidth function based on the bandwidth of all LED light sources in the LED multi-spectral imaging system;

[0076] Step S3, performing bandwidth iterative correction on the spectral reflectance to be corrected using the Richardson-Lucy iterative method according to the bandwidth function until a predetermined iteration end condition is met;

[0077] Step S4: After the iterative calculation is completed, the final correction bandwidth and step size are output as the optimal bandwidth correction parameters of the LED multi-spectral imaging system.

[0078] The specific process is as follows: before step S1, an LED multi-spectral imaging system is constructed, wherein the LED light source in the LED multi-spectral imaging system comes from an LED active lighting system.

[0079] The LED active lighting system is installed on the top of the lighting cabinet (with dimensions of 30 (D) × 40 (W) × 50 (H) cm). Figure 2 The LED active lighting system consists of 9 identical circuit boards, and M LED monochromatic light sources are evenly arranged on each circuit board.

[0080] The peak wavelengths of the M LEDs are evenly distributed in the range of 350nm to 780nm, and M is a positive integer greater than 1. In one example, the peak wavelengths of the light from the M=11 LED monochromatic light sources can be selected from 400nm, 425nm, 455nm, 475nm, 505nm, 525nm, 545nm, 605nm, 630nm, 670nm, and 700nm. The relative spectral power distribution of the LEDs is as follows: Figure 3 shown.

[0081] The camera is mounted in the center of the top of the lighting cabinet, below the LED active lighting system. A diffuser is mounted on the lens plane, and the lens is aimed at a known training color chart or predicted color chart placed horizontally at the bottom of the lighting cabinet. The diffuser is made of glass or acrylic.

[0082] In step S1, first, the standard reflectivity and spectral reflectivity of the predicted color card under the LED multi-spectral imaging system are obtained; secondly, the spectral reflectivity of the predicted color card is set as the spectral reflectivity to be corrected.

[0083] It should be noted that the spectral reflectance of the predicted color chart is obtained by performing the following steps:

[0084] S11. Select a training color chart and a prediction color chart. Both the training color chart and the prediction color chart are composed of color blocks containing multiple hues and chromaticities. For example, a 240-color color chart from Zhejiang Shike Co., Ltd. can be used; and an X-Rite ColorChecker Classic 24-color color chart from X-Rite Co., Ltd. can be used as the prediction color chart.

[0085] S12, using the LED multispectral imaging system to shoot, and obtain the camera response of each training color block in the training color card and the camera response of each predicted color block in the predicted color card;

[0086] S13. Obtaining the standard reflectance of each training color block and each predicted color block; wherein the standard reflectance can be obtained by an instrument such as a spectrophotometer; it should be noted that the standard reflectance of the predicted color block is used for result evaluation;

[0087] S14. Using an interpolation algorithm, interpolate the standard reflectance of each training color block and each predicted color block to the reflectance under the corresponding LED channel. It should be noted that since the wavelengths of the standard reflectance of the training color blocks are mostly distributed at intervals of 5nm or 10nm and cannot correspond one-to-one with the LED channels, the standard reflectance needs to be interpolated to the wavelength corresponding to the LED channel. The interpolation algorithm can be Lagrange fifth-order interpolation or cubic spline interpolation.

[0088] S15, establishing a conversion relationship between the camera response of the training color block and the standard reflectivity; wherein the conversion relationship is constructed using one or more methods of Wiener estimation method, pseudo-inverse method, and neural network;

[0089] S16. Calculate and obtain the spectral reflectance of the predicted color card based on the above conversion relationship.

[0090] In step S2, a bandwidth function is defined. The bandwidth function is constructed based on the full width at half maximum (FWHM) of the relative spectral power distribution of each LED light source. For example, the FWHMs of the 11 LEDs are approximately 17 nm, 19 nm, 22.5 nm, 26.5 nm, 29.5 nm, 32.5 nm, 39 nm, 18.5 nm, 19 nm, 23 nm, and 22 nm, respectively.

[0091] The bandwidth function is a Gaussian function (such as Figure 4 as shown) or a symmetrical triangular function; if the bandwidth function is a Gaussian function, the variance of each Gaussian function is obtained according to the full width at half maximum; if the bandwidth function is a symmetrical triangular function, the upper and lower limits of the abscissa of the piecewise function are obtained according to the full width at half maximum, so as to ensure that the area enclosed by the function and the x-axis is 1 before calculating the y-axis intercept.

[0092] In step S3, the steps of performing bandwidth iterative correction on the spectral reflectance to be corrected using the Richardson-Lucy iterative method specifically include:

[0093] (1.1) Initialization parameter m = m0. m0 can be 0.01;

[0094] (1.2) Initialization parameter n = n0, n0 can be 0.01;

[0095] (1.3) The line width of the bandwidth function under the same LED light source channel involved in the calculation is set to m times the FWHM of the initial test function, and the correction step size is set to n times the corresponding line width. The correction step size is δ = FWHM*m*n;

[0096] (1.4) Calculate the number of unilateral correction points N1 = [1 / n], N2 = [1 / n]. Where [·] represents the largest integer not greater than “·”, N1 represents the number of correction points to the left of the correction point, and N2 represents the number of correction points to the right of the correction point. Where N1 = N2, the total number of points in the bandwidth function is N = N1 + N2 + 1.

[0097] (1.5) The discretized bandwidth function is expressed as T stands for transpose;

[0098] (1.6) Calculate the first estimate of the measured spectrum:

[0099]

[0100] in, is the initial estimated spectrum, which is the reflectance of the spectrum to be corrected; Mirror image of bandwidth function; is the spectrum estimate at i sampling intervals (δ) to the right of the central wavelength point; * represents convolution; K is the number of points to be corrected, which is the number of LED types;

[0101] (1.7) Calculate the correction factor:

[0102]

[0103] Among them, M(λ k ) is the measured spectrum, i.e. the obtained spectral reflectance to be corrected;

[0104] (1.8) Convolve the correction factor with the bandwidth function to obtain the correction term:

[0105]

[0106] (1.9) The correction term is combined with the previous estimate Multiply them together to get a new estimate:

[0107]

[0108] The overall calculation process can also be expressed as follows:

[0109]

[0110] in is the updated value, is the value before update, r=0,1,2…, r+1 is the number of iterations;

[0111] (1.10) Return to step (1.6) until the set iteration end point is reached; the end point can be set to the difference between the two iterations before and after or the number of iterations, which can be set to or r+1≥100;

[0112] (1.11) Update the parameter n = n + e and return to step (1.3) until n increases to 1; where e can be 0.01;

[0113] (1.12) Update the parameter m = m + e and return to step (1.2) until m increases to 2; where e can be 0.01;

[0114] In step S4, after the iterative calculation is completed, the final correction bandwidth and step size are output as the optimal bandwidth correction parameters of the LED multi-spectral imaging system.

[0115] In an embodiment of the present invention, to obtain the optimal bandwidth for an LED multispectral imaging system, the initialization parameters m and n in the Richardson-Lucy iterative method can be set differently. The results of each iteration are compared to find the optimal solution, thereby adjusting the optimal values of parameters m and n. Therefore, the method further includes comparing the bandwidth correction results under different parameters with the standard reflectance of the predicted color chart to obtain the optimal parameters.

[0116] That is, first, in the Richardson-Lucy iterative method, the values of the initialization parameters m and n are adjusted multiple times;

[0117] Secondly, the bandwidth correction result obtained when the spectral reflectance to be corrected is subjected to bandwidth iterative correction using the Richardson-Lucy iterative method with each parameter adjustment is compared with the standard reflectance of the predicted color card to obtain the optimal values of m and n.

[0118] For example, (2.1) interpolate the corrected reflectance under each adjusted value of m and n to the same wavelength range of the spectrophotometer; wherein the interpolation algorithm uses Lagrange fifth-order interpolation or cubic spline interpolation;

[0119] (2.2) Compare the interpolated corrected reflectance with the standard reflectance of the predicted color block in the predicted color card and calculate the root mean square error; or, convert the interpolated corrected reflectance and the standard reflectance of the predicted color block into color values under a standard light source, and calculate the color difference between the corrected color and the standard color using a color difference formula; wherein the color difference formula can be the CIE76 color difference formula or the CIEDE2000 color difference formula.

[0120] (2.3) By comparing the root mean square error or color difference value under different values of m and n (such as Figure 5 As shown in FIG, the optimal m and n values are obtained as the optimal bandwidth correction parameter combination of the LED multispectral imaging system.

[0121] like Figure 6 As shown in the figure, an LED multi-spectral imaging system bandwidth correction system is provided in an embodiment of the present invention, comprising:

[0122] The spectral reflectance acquisition unit 110 is used to acquire the standard reflectance and spectral reflectance of the predicted color card under the LED multi-spectral imaging system, and set the spectral reflectance of the predicted color card as the spectral reflectance to be corrected;

[0123] A bandwidth function definition unit 120 is configured to define a bandwidth function based on the bandwidths of all LED light sources in the LED multi-spectral imaging system;

[0124] a bandwidth correction iteration unit 130 for performing bandwidth iterative correction on the spectral reflectance to be corrected according to the bandwidth function using a Richardson-Lucy iteration method until a predetermined iteration end condition is met;

[0125] The correction bandwidth output unit 140 is used to output the final correction bandwidth and step size obtained after the iterative calculation is completed as the optimal bandwidth correction parameters of the LED multi-spectral imaging system.

[0126] Wherein, the LED light source in the LED multi-spectral imaging system comes from the LED active lighting system; wherein,

[0127] The LED active lighting system is installed on top of the pre-installed lighting cabinet and above the camera. It consists of 9 identical circuit boards, and each circuit board has M LED monochromatic light sources evenly arranged; the peak wavelengths of the M LEDs are evenly distributed in the range of 350nm to 780nm;

[0128] The camera is installed in the middle of the top of the lighting cabinet and below the LED active lighting system. A light-diffusing plate is installed on the lens plane, and the lens is aimed at a known training color card or a predicted color card placed horizontally at the bottom of the lighting cabinet.

[0129] The bandwidth function is constructed based on the full width at half maximum obtained from the relative spectral power distribution of each LED light source;

[0130] The bandwidth function is a Gaussian function or a symmetrical triangular function; if the bandwidth function is a Gaussian function, the variance of each Gaussian function is obtained according to the full width at half maximum; if the bandwidth function is a symmetrical triangular function, the upper and lower limits of the horizontal coordinate of the piecewise function are obtained according to the full width at half maximum.

[0131] The implementation of the embodiments of the present invention has the following beneficial effects:

[0132] The present invention defines a bandwidth function based on the bandwidth of all LED light sources in the LED multispectral imaging system. The Richardson-Lucy iterative method is used to perform bandwidth iterative correction on the spectral reflectance to be corrected, and the optimal correction parameters of the LED multispectral imaging system are output, thereby solving the problem that the accuracy of spectral reflectance reconstruction of the LED multispectral imaging system is difficult to further improve.

[0133] It is worth noting that in the above system embodiment, the various units included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0134] Those skilled in the art will understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc.

[0135] The above disclosure is only a preferred embodiment of the present invention and certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for bandwidth correction of LED multispectral imaging system, characterized in that: The method comprises the following steps: Obtaining the standard reflectance and spectral reflectance of the predicted color card under the LED multispectral imaging system, and setting the spectral reflectance of the predicted color card as the spectral reflectance to be corrected; Based on the bandwidth of all LED light sources in the LED multispectral imaging system, a bandwidth function is defined; According to the bandwidth function, using the Richardson-Lucy iterative method, performing bandwidth iterative correction on the spectral reflectance to be corrected until a predetermined iteration end condition is met; After the iterative calculation is completed, the final correction bandwidth and step size are output as the optimal bandwidth correction parameters of the LED multi-spectral imaging system; The bandwidth function is constructed based on the full width at half maximum obtained from the relative spectral power distribution of each LED light source; wherein, The bandwidth function is a Gaussian function or a symmetrical triangular function; if the bandwidth function is a Gaussian function, the variance of each Gaussian function is obtained according to the full width at half maximum; if the bandwidth function is a symmetrical triangular function, the upper and lower limits of the horizontal coordinate of the piecewise function are obtained according to the full width at half maximum.

2. The LED multi-spectral imaging system bandwidth correction method according to claim 1, wherein: The LED light source in the LED multi-spectral imaging system comes from the LED active lighting system; wherein, The LED active lighting system is installed on the top of the pre-set lighting cabinet and is located above the camera. It consists of 9 identical circuit boards, and each circuit board has LED monochromatic light sources are evenly arranged; The peak wavelengths of the LEDs are evenly distributed in the range of 350nm~780nm; The camera is installed in the middle of the top of the lighting cabinet and below the LED active lighting system. A light-diffusing plate is installed on the lens plane, and the lens is aimed at a known training color card or a predicted color card placed horizontally at the bottom of the lighting cabinet.

3. The LED multi-spectral imaging system bandwidth correction method according to claim 1, characterized in that: The step of performing bandwidth iterative correction on the spectral reflectance to be corrected using the Richardson-Lucy iterative method according to the bandwidth function until a predetermined iteration end condition is met specifically includes: (1.1) Initialization parameter m=m0, m0 can be 0.01; (1.2) The initialization parameter n=n0, n0 can be 0.01; (1.3) Set the line width of the bandwidth function under the same LED light source channel to the FWHM of the initial test function. m times, the correction step is set to the corresponding line width n times, the correction step size is ; (1.4) Calculate the number of unilateral correction points: N1 = [1 / n], N2 = [1 / n], where [·] represents the largest integer not greater than "·", N1 represents the number of correction points to the left of the correction point, and N2 represents the number of correction points to the right of the correction point. Where N1 = N2, the total number of points in the bandwidth function is N = N1 + N2 + 1. (1.5) The discretized bandwidth function is expressed as , T represents transpose; (1.6) Calculate the first estimate of the measured spectrum: in, is the initial estimated spectrum, which is the reflectance of the spectrum to be corrected; Mirror image of bandwidth function; To the right from the center wavelength point i Sampling intervals ( )’s spectral estimate; represents convolution; K The number of points to be calibrated is the number of LED types; (1.7) Calculate the correction factor: in, The measured spectrum, i.e. the obtained spectral reflectance to be corrected; (1.8) Convolve the correction factor with the bandwidth function to obtain the correction term: ； (1.9) The correction term is added to the previous estimate. Multiply them together to get a new estimate: ； The overall calculation process can also be expressed as follows: in is the updated value, is the value before updating, r= 0,1,2…, r +1 is the number of iterations; (1.10) Return to step (1.6) until the set iteration end point is reached; the end point can be set to the difference between the two iterations before and after or the number of iterations, which can be set to or ; (1.11) Update the parameter n = n + e and return to step (1.3) until n increases to 1; where e can be 0.01; (1.12) Update the parameter m = m + e and return to step (1.2) until m increases to 2; where e can be 0.01; 4. The LED multi-spectral imaging system bandwidth correction method according to claim 3, wherein: The method further comprises: In the Richardson-Lucy iterative method, the values of the initialization parameters m and n are adjusted multiple times; The bandwidth correction result obtained when the spectral reflectance to be corrected is subjected to bandwidth iterative correction using the Richardson-Lucy iterative method with each parameter adjustment is compared with the standard reflectance of the predicted color card to obtain the optimal values of m and n.

5. The LED multi-spectral imaging system bandwidth correction method according to claim 4, characterized in that: The step of comparing the bandwidth correction result obtained by performing bandwidth iterative correction on the spectral reflectance to be corrected using the Richardson-Lucy iterative method with each parameter adjustment with the standard reflectance of the predicted color card to obtain the optimal values of m and n specifically includes: Interpolating the corrected reflectance under each adjusted value of m and n to the same wavelength range of the spectrophotometer; wherein the interpolation algorithm uses Lagrange fifth-order interpolation or cubic spline interpolation; Comparing the interpolated corrected reflectivity with the standard reflectivity of the predicted color block in the predicted color card to calculate a root mean square error value; or converting the interpolated corrected reflectivity and the standard reflectivity of the predicted color block into color values under a standard light source, and calculating the color difference between the corrected color and the standard color using a color difference formula; wherein the color difference formula can be a CIE76 color difference formula or a CIEDE2000 color difference formula; By comparing the root mean square error or color difference value under different values of m and n, the optimal m and n values are obtained as the optimal bandwidth correction parameter combination of the LED multi-spectral imaging system.

6. The LED multi-spectral imaging system bandwidth correction method according to claim 2, wherein: The spectral reflectance of the predicted color card is obtained by performing the following steps: Selecting a training color card and a prediction color card; wherein the training color card and the prediction color card are both composed of color blocks containing multiple hues and chromas; Using the LED multispectral imaging system to shoot, obtain the camera response of each training color block in the training color chart and the camera response of each predicted color block in the predicted color chart; Obtaining a standard reflectance of each training color block and each predicted color block; wherein the standard reflectance can be obtained by an instrument such as a spectrophotometer; Use the interpolation algorithm to interpolate the standard reflectance of each training color block and each predicted color block to the reflectance under the corresponding LED channel; Establishing a conversion relationship between the camera response of the training color block and the standard reflectivity; wherein the conversion relationship is constructed using one or more methods of Wiener estimation method, pseudo-inverse method, and neural network; The spectral reflectance of the predicted color card is calculated based on the conversion relationship.

7. A bandwidth correction system for LED multi-spectral imaging system, characterized in that: include; The spectral reflectance acquisition unit to be corrected is used to obtain the standard reflectance and spectral reflectance of the predicted color card under the LED multi-spectral imaging system, and set the spectral reflectance of the predicted color card as the spectral reflectance to be corrected; A bandwidth function definition unit, configured to define a bandwidth function based on the bandwidths of all LED light sources in the LED multi-spectral imaging system; a bandwidth correction iteration unit, configured to perform bandwidth iterative correction on the spectral reflectance to be corrected according to the bandwidth function using a Richardson-Lucy iteration method until a predetermined iteration end condition is met; A correction bandwidth output unit is used to output the final correction bandwidth and step size obtained after the iterative calculation is completed as the optimal bandwidth correction parameters of the LED multi-spectral imaging system; The bandwidth function is constructed based on the full width at half maximum obtained from the relative spectral power distribution of each LED light source; wherein, The bandwidth function is a Gaussian function or a symmetrical triangular function; if the bandwidth function is a Gaussian function, the variance of each Gaussian function is obtained according to the full width at half maximum; if the bandwidth function is a symmetrical triangular function, the upper and lower limits of the horizontal coordinate of the piecewise function are obtained according to the full width at half maximum.

8. The LED multi-spectral imaging system bandwidth correction system according to claim 7, characterized in that: The LED light source in the LED multi-spectral imaging system comes from the LED active lighting system; wherein, The LED active lighting system is installed on the top of the pre-set lighting cabinet and is located above the camera. It consists of 9 identical circuit boards, and each circuit board has LED monochromatic light sources are evenly arranged; The peak wavelengths of the LEDs are evenly distributed in the range of 350nm~780nm; The camera is installed in the middle of the top of the lighting cabinet and below the LED active lighting system. A light-diffusing plate is installed on the lens plane, and the lens is aimed at a known training color card or a predicted color card placed horizontally at the bottom of the lighting cabinet.

Images