Lamp bead light intensity control method for maximizing finger vein contrast under non-uniform illumination

Abstract

The invention relates to the technical field of living body finger vein recognition, in particular to a lamp bead light intensity control method for maximizing the finger vein contrast ratio under non-uniform illumination. The method comprises the following steps: S1, before the contrast of a finger vein image under non-uniform illumination is maximized, each acquisition device only makes preparation work once; s2, when the finger vein images are collected again, (k + 1) groups of finger vein images with different light intensities are shot again, and according to the light intensity superposition principle, gray values # imgabs0 # and # imgabs1 # under driving of any combination of duty ratios are calculated; S3, the duty ratios pk of driving PWM waves of K LED light intensities are found by adopting a genetic algorithm when the sum of column pixel gray contrast ratios is maximum. According to the lamp bead intensity control method for maximizing the finger vein contrast ratio under the non-uniform illumination, by controlling the lamp bead intensity of each lamp bead in the near-infrared light source, the gray scale contrast ratio sum value of pixel columns in a finger contour in an image obtained after transmission finger vein radiography is maximum.

Description

Technical Field

[0001] The present invention relates to the technical field of living finger vein recognition, and in particular to a method for controlling the light intensity of lamp beads to maximize the contrast of finger veins under non-uniform illumination. Background Art

[0002] Currently, finger vein recognition technology has become a hot area in biometric recognition technology. The grayscale contrast of the collected finger vein image has a great impact on the subsequent vein pattern enhancement, binarization and feature extraction / recognition, so it is necessary to maximize the contrast of the finger vein image during acquisition.

[0003] The existing near-infrared light source uses several LEDs to form a lamp array. When the user's finger is placed in the collector, the lamp array will first emit a combined light intensity to transmit the finger vein imaging, quickly collect the finger vein image under the current light intensity for some image quality analysis, and then adjust the light intensity of each lamp according to the analysis results. This process is a negative feedback adjustment method. Generally, the adjustment of the lamp intensity will be stopped after the quality of the obtained vein image meets the standard. The grayscale range of this indicator is relatively wide, mainly to avoid overexposure and underexposure, but it is still prone to low contrast. The vein grayscale distribution of the same finger may be quite different when the vein images are collected at different times, that is, the intra-class distance is large. This requires the later software to perform contrast enhancement and grayscale normalization, which will increase the difficulty and running time of the algorithm, and is not conducive to the real-time application of vein recognition.

[0004] To this end, a method for controlling the light intensity of a lamp bead is designed to maximize the contrast of finger veins under non-uniform illumination, so as to provide another technical solution to the above technical problems. Summary of the Invention

[0005] Based on this, it is necessary to provide a lamp intensity control method for maximizing the contrast of finger veins under non-uniform illumination to address the above technical problems, so as to solve the technical problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] The steps for controlling the light intensity of the LEDs to maximize the contrast of finger veins under non-uniform illumination are as follows:

[0008] S1: Each acquisition device was prepared only once before maximizing the contrast of finger vein images under non-uniform illumination;

[0009] The preparation work is used to calculate the camera response function CRF of the corresponding acquisition device;

[0010] During the acquisition process, a finger is selected and each light bead is illuminated multiple times individually or in combination to collect a multi-intensity finger vein image sequence;

[0011] S2: When re-collecting the finger vein image, re-shoot k+1 groups of finger vein images with different light intensities, and calculate the grayscale value under any combination of duty cycle driving according to the principle of light intensity superposition and

[0012] S3: Use genetic algorithm to find the duty cycle p of the driving PWM wave of K LED light intensity when the sum of the column pixel grayscale contrast is the maximum value k .

[0013] As a preferred embodiment of the method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination provided by the present invention, in step S1, a finger is selected, and each lamp bead is illuminated multiple times individually or in combination to capture a multi-intensity finger vein image sequence, as follows:

[0014] S11: By individually controlling the brightness of the lamp beads, all lamp beads are driven to illuminate a finger vein imaging with corresponding brightness;

[0015] S12: Calculate the camera response function, which is expressed as follows:

[0016] I=f(E)

[0017] Where f is the CRF of each photoelectric CMOS sensor; the input of function f is the light intensity E, and the output is the pixel grayscale I;

[0018] S13: Estimate the CRF of the camera and the approximate duty cycle range of the tentative light intensity. The steps are as follows:

[0019] S131: After extracting the appropriate region of interest, the Debevec method is used to estimate the camera's CRF;

[0020] S132: Based on the approximate duty cycle range of the tentative light intensity and the K lamp beads of the near-infrared light source lamp bead array, multiple combined lighting of K lamp beads is performed to ensure that there is no overexposure or underexposure in K+1 combined lightings, and the light intensity with the minimum duty cycle is listed as the known reference light intensity.

[0021] As a preferred embodiment of the lamp bead light intensity control method for maximizing the finger vein contrast under non-uniform illumination provided by the present invention, in step S132, the ratio of the light intensity with the minimum duty cycle to the duty cycles of other light intensities is r.

[0022] As a preferred embodiment of the lamp intensity control method for maximizing the contrast of finger veins under non-uniform illumination provided by the present invention, in step S2, when re-collecting the finger vein image, k+1 groups of finger vein images with different light intensities are re-taken, and the grayscale value under any combination of duty cycle driving is calculated according to the principle of light intensity superposition. and Here are the steps:

[0023] S21: According to the total light intensity E of each column of fingers j The pixel grayscale obtained by angiography is used to obtain the light intensity response equations of multiple K lamp combinations;

[0024] S22: Solve the equations to find the minimum duty cycle of each lamp bead and the maximum grayscale value A of each column in the finger vein image. j and grayscale minimum value B j IRC offset value and

[0025] S23: Calculate the two points A in each column j and B j Grayscale value under the duty cycle drive of any combination of K lamp beads and

[0026] As a preferred embodiment of the method for controlling the light intensity of the lamp beads to maximize the contrast of the finger veins under non-uniform illumination provided by the present invention, the total light intensity E of each column of the finger is j It is K lamp beads with a light intensity of E jk The superposition expression is as follows:

[0027]

[0028] All pixels I(i,j) (i=1,2,…,M) in the jth column are illuminated by the uniform light intensity E of the kth lamp bead. jk When the contrast is in effect, the pixel grayscale obtained is expressed as follows:

[0029] I(i,j,k)=f(E ijk +T i,j )j=1,2,…,N,k=1,2,…,K,i=1,2,…,M.

[0030] As a preferred embodiment of the method for controlling the light intensity of the lamp beads to maximize the contrast of the finger vein under non-uniform illumination provided by the present invention, in step S23, the two points A in each column are calculated. j and B j Grayscale value under the duty cycle drive of any combination of K lamp beads and Here are the steps:

[0031] S231: Grayscale value and The expression is as follows:

[0032]

[0033] Two Points A j and B j The grayscale difference is expressed as follows:

[0034]

[0035] According to j = 1, 2, ..., N, and combined with the overall contrast of the finger vein image, the sum of multiple column contrasts C and The expression is as follows:

[0036]

[0037] in,

[0038] As a preferred embodiment of the method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination provided by the present invention, in step S231, the overall contrast of the finger vein image is the sum C of the contrasts of multiple columns, which is expressed as follows:

[0039]

[0040] As a preferred embodiment of the method for controlling the light intensity of LEDs to maximize the contrast of finger veins under non-uniform illumination provided by the present invention, in step S3: a genetic algorithm is used to find the K LED light intensity combinations that achieve the maximum value, and the steps are as follows:

[0041] According to the light intensity of K lamp beads in each column and the IRC offset of points A and B in each column of the finger image and

[0042] The genetic algorithm is used to calculate the best combination and obtain the pk when the column grayscale contrast and value are the maximum;

[0043] The actual circuit is driven to complete the finger vein image when the sum of the grayscale contrast of all pixel columns reaches the maximum value. k .

[0044] It can be seen without a doubt that the above-mentioned technical solution of this application can definitely solve the technical problem to be solved by this application.

[0045] At the same time, through the above technical solutions, the present invention has at least the following beneficial effects:

[0046] 1. The present invention provides a method for controlling the light intensity of lamp beads to maximize the contrast of finger veins under non-uniform illumination. By controlling the light intensity of each lamp bead in the near-infrared light source, the grayscale contrast and value of the pixel columns in the finger contour in the image after transmission finger vein angiography are maximized.

[0047] 2. The present invention uses the known camera response function CRF to model the illumination response curve of the pixel column, combined with the known column light intensity distribution when the lamp beads are illuminated, and obtains the pixel grayscale contrast of each column through the superposition principle. Finally, a genetic algorithm is used to solve the maximum sum of the pixel column grayscale contrast, thereby maximizing the contrast when acquiring finger vein images under the condition of hardware-adjusted non-uniform light source distribution. This is beneficial to improving the overall grayscale contrast during transmission-type finger vein angiography imaging acquisition, can reduce the intra-class distance, and make the grayscale distribution of the same finger more similar during multiple acquisitions, which is beneficial to subsequent applications such as vein segmentation, feature extraction, and identity recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 This is a schematic diagram of the light intensity distribution on multiple layered planes during transmission finger vein imaging using a multi-lamp array in the prior art;

[0050] Figure 2 Schematic diagram of the CRF estimated from the multi-intensity finger vein image sequence of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0052] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0053] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features and technical solutions therein may be combined with each other.

[0054] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0055] Example 1

[0056] refer to Figure 1 , a method for controlling the light intensity of lamp beads to maximize the contrast of finger veins under non-uniform illumination.

[0057] Preliminary research has shown that when the maximum grayscale difference of pixel points in the finger body is defined as contrast, the camera illumination response function (CRF) can be used to model the image when uniform light is transmitted through the finger vein. The illumination response curve (IRC) of each pixel is a translation of the CRF, and the light intensity that maximizes the grayscale contrast can be calculated. However, among existing finger vein devices, most use a lamp array, in which multiple lamps are arranged in a straight line with equal spacing. The illumination distribution of this type of near-infrared light source on the dorsal plane of the finger (hereinafter referred to as the dorsal plane) is relatively complex, as shown below. Figure 1 As shown, the light intensity of each lamp bead is the brightest at the vertical projection point. The farther away from the projection point, the more the light intensity decays. The light intensity distribution between each lamp bead will overlap, as shown below Figure 1 In plane 1, the direct modeling of the light intensity distribution of multiple lamp beads superimposed on this plane is relatively complex. In addition, the diffusion effect of the physiological tissue of the finger body causes the light beam that penetrates the finger to be scattered, absorbed, and diffusely reflected before reaching the superficial subcutaneous vein plane on the inside of the finger (hereinafter referred to as the vein plane). The light intensity distribution can be visualized as follows Figure 1 In plane 2, the light intensity of each column in the direction perpendicular to the finger is uniform, and then after venography, the imaging is captured by the camera.

[0058] In the early experiments, it was found that when the light of the lamp beads penetrates the finger, the light intensity will be uniform in the vertical direction of the venography surface due to the absorption and diffusion effect of the finger's physiological tissue, but there will still be intensity changes in the horizontal row (such as Figure 1 The light intensity distribution of plane 2 in the image is obtained by calculating the intensity distribution of multiple lamp beads, and then the light intensity superposition principle during transmission imaging can be used to model the light intensity of multiple lamp beads, establish an overdetermined set of equations for the optimal light intensity distribution, and solve a relatively optimal solution. When registering and recognizing finger vein images, the light intensity distribution of each lamp bead can be obtained by testing the light intensity of several groups of light source lamp beads. Combining CRF and optimization methods, the relatively optimal solution can be calculated to improve the overall contrast of the collected finger vein images.

[0059] In this embodiment, Figure 1In the figure, Plane 1 represents the surface of the finger where light from the light source reaches the back of the finger. The upper half of the finger, which is approximately cylindrical, is illuminated. Plane 2 represents the surface of the finger where the near-infrared light passes through the finger, is scattered, reflected, and absorbed within the finger, and reaches the lower half of the finger, which is approximately cylindrical and close to the skin (before imaging of the superficial subcutaneous venous patterns). Plane 3 represents the surface where the superficial subcutaneous venous patterns are located. The names Plane 1, Plane 2, and Plane 3 are provided for ease of understanding.

[0060] 1. The settings are as follows:

[0061] The present invention maximizes the grayscale contrast of the collected finger venography image by controlling the PWM wave duty cycle of multiple lamp beads. The principle and settings are as follows:

[0062] (i) According to references [A] and [B], the CRF of each photoelectric CMOS sensor in a camera is identical, denoted as f, and can be calculated in advance. This characteristic is stable and unchanging, requiring only a single calculation for subsequent use. This function is monotonically increasing and has an inverse function g, which is also monotonically increasing. The input of this function f is the light intensity E, and the output is the pixel grayscale I. The grayscale is typically quantized to 8 bits deep, expressed as follows:

[0063] I=f(E) (1)

[0064] Among them, the document [A] is PEDebevec, J.Malik.Recovering high dynamic rangeradiance maps from photographs[M] / / Seminal Graphics Papers: Pushing the Boundaries, 2023, 2: 643-652.

[0065] The document [B] is:]PEDebevec and J.Malik,Recovering high dynamic rangeradiance maps from photographs[C] / / Proceedings of the 24th annual conferenceon Computer graphics and interactive techniques.1999:369-378

[0066] Among them, document [A] is a monograph after document [B], which provides some methods for correcting the camera response function. Figure 2It is concluded that the multi-exposure illumination response curves of multiple points on the image are the translations of the camera response function curves.

[0067] (ii) According to references [A] and [B], when the light intensity is changed for acquisition, the IRC of each pixel (i, j) is the translation of the CRF curve on the light intensity axis. During finger venography, the IRC translation T(i, j) varies depending on the absorption / transmission intensity of the finger position corresponding to each pixel position. This characteristic objectively reflects the near-infrared absorption characteristics of each point on the finger during imaging, and is expressed as follows:

[0068] I(i,j)=f(E(i,j)+T(i,j)) (2)

[0069] (iii) The near-infrared light source array has K lamp beads, and the light intensity of each lamp bead can be expressed by the PWM wave duty cycle p k Control, its light intensity is proportional to the duty cycle, when the kth lamp is lit alone, the light intensity is recorded as E k , the expression is as follows:

[0070]

[0071] All pixels I(i,j) (i=1,2,...,M) in the jth column are illuminated by the uniform light intensity E of the kth lamp bead. jk When the contrast is applied, the pixel grayscale obtained by the imaging is recorded as the following formula (4). The coordinates of the light intensity and IRC offset value are marked in the lower right corner. The expression is as follows:

[0072] I(i,j,k)=f(E ijk +T i,j )j=1,2,…,N,k=1,2,…,K,i=1,2,…,M (4)

[0073] (iv) The absorption and diffusion effect of the finger physiological tissue forms a uniform light intensity in the vertical column direction of the vein plane. The light intensity of each column in the finger vein image is uniform. The footer of the light intensity E in formula (4) is denoted as ijk, and there is formula (5):

[0074] E 1jk =E 2jk =…=E Mjk j=1,2,…,N k=1,2,…,K (5)

[0075] Simplify i and write it as E jk , the kth lamp bead has a duty cycle p q When the driver is on, the uniform light intensity transmitted / scattered to the jth column is recorded as The lower right corner of k is marked with p q Indicates that the duty cycle of the kth lamp bead is pq , when the duty cycle of the lamp changes to r·p q When r is the variation factor, the uniform light intensity of the jth column is recorded as Then from formula (3), we get the following light intensity change relationship:

[0076]

[0077] During the finger vein image acquisition process, trial photos are taken with light intensities of several duty cycles. After taking the photos, light intensity adjustment is performed based on image analysis under these light intensities. This provides a light intensity analysis basis for subsequent light intensity adjustment and grayscale contrast estimation.

[0078] (v) The local contrast of the collected finger vein image (the finger points from left to right) is defined by pixel column. Assuming that the size of the finger vein image I is M×N, the grayscale contrast of each column can be defined as the maximum minus the minimum grayscale value of the pixels in the column.

[0079] C j =max(I(x,j))-min (I(x,j)) j=1,2,…,N (7)

[0080] The overall contrast of a finger vein image is the sum of the contrasts of multiple columns:

[0081]

[0082] The task of the present invention is to find the following formula (9):

[0083]

[0084] That is, the K lamp duty cycle values p1, p2, ..., pK when the sum of all column contrasts in image I is the largest.

[0085] (vi) According to the literature [C], when multiple LEDs are turned on, the finger vein image obtained is the superposition of images of each LED with the same brightness turned on individually. The k1th LED is turned on individually, and its duty cycle is recorded as p1. The finger vein image obtained is recorded as Similarly, the k2th lamp is turned on alone, and its duty cycle is recorded as p2. The finger vein image obtained is recorded as When K1 and K2 lamps are on at the same time, and the duty cycles are p1 and p2 respectively, the finger vein image obtained is recorded as According to the superposition principle, we have the following formula (10):

[0086]

[0087] Among them, the document [C] is [Bichsel M, Ohnesorge K W. How to Measure a Camera'sResponse Curve from Scratch[J]. University of Zurich, 1993].

[0088] In the jth column pixel of the finger vein image, under uniform illumination conditions (Formula 5), if A j Point B is the maximum gray value of the column, and j Point 1 is the grayscale minimum value of the column. When the illumination intensity changes, these two points always maintain the maximum and minimum values of the column. In other words, the local grayscale contrast of the column (Formula 7) is determined by these two points. The specific calculation method is provided in the authorized invention patent (CN114677712A - Method for maximizing local contrast of finger vein images under uniform illumination);

[0089] The present invention refers to this method, and only needs to estimate the grayscale difference between these two points to be the local contrast of the column. j Point and B j The light intensity with the largest grayscale difference

[0090] However, in actual illumination, due to the total light intensity E of each column of fingers j It is K lamp beads with a light intensity of E jk Overlay:

[0091]

[0092] When formula (12) is equal to formula (11), there is the optimal solution of formula (9), namely:

[0093]

[0094] Formula (13) is a system of equations. There are N columns of optimal light intensity values (constraints) on the left, but only K variables on the right, and N>>K. This is an overdetermined equation and may not have an optimal solution. If there is no optimal solution in the present invention, a second-best solution can be sought, which satisfies Formula (9).

[0095] 2. The calculation is as follows:

[0096] The method of the present invention is divided into two parts. The first part only needs to be done once, which is equivalent to calibrating the acquisition equipment. The second part is to dim the finger veins to maximize the contrast.

[0097] Part 1: First, select a finger (a prosthetic finger is also acceptable), collect a multi-intensity finger vein image sequence, and calculate the camera response function (1). Then, light up each LED individually or in combination multiple times, collect a multi-intensity finger vein image sequence, and estimate the camera's CRF and the approximate duty cycle range of the tentative light intensity. The tentative light intensity refers to the light intensity that can ensure that the finger vein of a general thickness is not under-exposed or over-exposed when imaging.

[0098] Part II: The process of maximizing the grayscale contrast during acquisition is as follows. Based on the approximate range of the tentative light intensity determined in the first part, multiple K lamp beads are combined for illumination to ensure that there is no overexposure or underexposure in the K+1 combined illuminations. The light intensity column with the minimum duty cycle can be used as the known reference light intensity. The ratio r of the duty cycles of other light intensities to this light intensity duty cycle is known, as shown in formula (6). Then, using the principle of light intensity superposition, the light intensity response equations for multiple K lamp beads are obtained from formulas (12) and (4). The equations can be solved to obtain the minimum duty cycle of each lamp bead and the maximum grayscale value A of each column in the finger vein image. j and grayscale minimum value B j IRC offset value and It is also possible to estimate the two points A in each column j and B j Grayscale value under the duty cycle drive of any combination of K lamp beads and

[0099]

[0100] Formula (7) is transformed into two points A j and B j Grayscale difference:

[0101]

[0102] In formula (15), j = 1, 2, ..., N. Combining formula (8) and formula (9), we can get:

[0103]

[0104] Obviously, this formula has a maximum value, but it is almost impossible to reach the optimal light intensity value (Formula 13) of the maximum grayscale contrast of each pixel column. In order to find the maximum value of Formula (16), the present invention uses a genetic algorithm to find the LED light intensity combination that reaches the maximum value.

[0105] The genetic algorithm is used to search for the maximum value of formula (16), which avoids multiple attempts to solve the combination of K lamp beads. Each lamp bead has 256 light intensities, and there are 256 in total. KIn practical applications, it is impossible to collect all the combinations. After trying to turn on Q lamp beads Q+1 times, the present invention can calculate the light intensity of K lamp beads in each column. and the IRC offset of points A and B in each column of the finger image and Then use the genetic algorithm to find the best combination and obtain the maximum value of the column grayscale contrast and p k .

[0106] Example 2

[0107] Reference Figure 2 , based on the above embodiment 1, an implementation method thereof is disclosed.

[0108] Taking a vein collection platform with an LED array light source as an example, the light source array has 5 lamp beads, each of which can control the brightness individually and adjust the duty cycle of the PWM wave to adjust the light intensity and brightness. First, all the lamp beads are driven to light up a finger vein imaging with the same brightness. The brightness is The collected images are denoted as I1, I2,…, I 26 ,After taking out the appropriate region of interest (Region of Interest ROI), the Debevec method is used to estimate the CRF of the camera, such as Figure 2 ;

[0109] Among them, the region of interest is the part that needs to be studied and processed, which is generally a rectangle.

[0110] When applying, it is necessary to collect a registration or authentication finger vein image. In this process, all five lamp beads are turned on with the same duty cycle p1, and the vein image obtained is recorded as Here, because they are all the same duty cycle, it is still simply recorded as The grayscale maximum point of the jth column can be found, denoted as A j Point, its gray value is According to the superposition principle, we can get:

[0111]

[0112] here It represents the light intensity acting on the jth column of the finger vein image when the PWM duty cycle of LED light number k is p1. A j The offset of the illumination response curve of the point relative to the CRF, f is the CRF function.

[0113] Then set the duty cycle p2=r2*p1, r2>1, so that each LED light is lit with the duty cycle of p2, and the gray value of point A is We can get:

[0114]

[0115] Obviously, when all LED lights are turned on for the second time, the light intensity of the jth column is increased by r2 times according to the previous formula (6). Formula (18) is converted to formula (19):

[0116]

[0117] The duty cycle is changed 4 times in the same way as above. The growth rates of the duty cycle relative to the first duty cycle p1 are r3, r4, r5, and r6 respectively. Then A is collected. j The gray value of the point, we can get another four equations, such as equations (20) to (23):

[0118]

[0119] The combined equations (17), (19), (20), (21), (22) and (23) form the equation system, i.e., A in the jth column j The pixel corresponds to a 6-order illumination response equation group with 6 unknowns, which are The equations are solvable (without underexposure and overexposure, I A ≠0 and I A ≠255), solve these 6 unknowns, and get the light intensity of each LED light combined in the jth column at duty cycle p1. The light intensity offset value of the minimum gray value point B in the jth column can be obtained in the same way.

[0120] Repeat the above process to solve each column

[0121] Use the same method to obtain other columns and And the light intensity distribution E of the kth LED lamp in the jth column jk , j = 1, 2, ..., N, k = 1, 2, ..., 5, with these distribution parameters and CRF, the genetic algorithm can be used to solve equation (19) as follows:

[0122] (a) Individual coding: Encode the brightness of the column directly below each LED bead, which is represented by chromosome E in the genetic algorithm. Since the PWM wave range of each light is 0 to 255, it can be encoded using 8 bits. The expression is as follows:

[0123] E=[E1,E2,…,E K ] (twenty four)

[0124] (b) Initialize the population and randomly generate D groups of initial solutions E1 ,E 2 ,…,E D ;

[0125] (c) Fitness function. The fitness function of each individual in the population is defined as formula (25), as follows:

[0126]

[0127] here is the light intensity of the kth LED lamp acting on the jth column, and its duty cycle is p k , and the previous solution is The duty cycle of the two is p1, and the ratio of the two is recorded as Substituting into formula (25) we get formula (26), as follows:

[0128]

[0129] The equation (26) can be solved by referring to equations (17), (19), (20), (21), (22) and (23). Numerical calculation is performed, that is, the fitness value can be calculated. The larger the fitness, the larger the sum of the grayscale contrast of the pixel column, and the better the overall contrast.

[0130] (d) Selection operation: select individuals with higher fitness for reproduction, and methods such as roulette wheel selection and tournament selection can be used.

[0131] (e) Crossover operation: Use crossover operation to combine two individuals E d1 ,E d2 The bit values of the vector part of (d1,d2∈{1,2,…,D}) are crossed to generate a new individual.

[0132] (f) Mutation operation: some E in the individual s The components are mutated to ensure the diversity of the population.

[0133] (g) Update the population: replace old individuals with newly generated individuals and enter the next generation.

[0134] (h) Termination condition: When the maximum number of iterations is reached or the fitness is no longer significantly improved, the algorithm is terminated and the light intensity solution that maximizes equation (16) is output (the PWM duty cycle p of each lamp bead is obtained). k ).

[0135] The duty cycle p of these five lamp beads is calculated by genetic algorithm. k After that, the actual circuit can be driven to complete the p value when the sum of the grayscale contrast of all pixel columns of the finger vein image reaches the maximum value. k .

[0136] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. The preferred embodiments do not describe all details in detail, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination, characterized in that: Here are the steps: S1: Each acquisition device was prepared only once before maximizing the contrast of finger vein images under non-uniform illumination; The preparation work is used to calculate the camera response function CRF of the corresponding acquisition device; During the acquisition process, a finger is selected and each light bead is illuminated multiple times individually or in combination to collect a multi-intensity finger vein image sequence; S2: When re-collecting the finger vein image, re-shoot k+1 groups of finger vein images with different light intensities, and calculate the grayscale value under any combination of duty cycle driving according to the principle of light intensity superposition and S3: Use genetic algorithm to find the duty cycle p of the driving PWM wave of K LED light intensity when the sum of the column pixel grayscale contrast is the maximum value k .

2. The method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination according to claim 1, characterized in that: In step S1, a finger is selected and each LED is illuminated multiple times individually or in combination to collect a multi-intensity finger vein image sequence. The steps are as follows: S11: By individually controlling the brightness of the lamp beads, all lamp beads are driven to illuminate a finger vein imaging with corresponding brightness; S12: Calculate the camera response function, which is expressed as follows: I=f(E) Where f is the CRF of each photoelectric CMOS sensor; the input of function f is the light intensity E, and the output is the pixel grayscale I; S13: Estimate the CRF of the camera and the approximate duty cycle range of the tentative light intensity. The steps are as follows: S131: After extracting the appropriate region of interest, the Debevec method is used to estimate the camera's CRF; S132: Based on the approximate duty cycle range of the tentative light intensity and the K lamp beads of the near-infrared light source lamp bead array, multiple combined lighting of K lamp beads is performed to ensure that there is no overexposure or underexposure in K+1 combined lightings, and the light intensity with the minimum duty cycle is listed as the known reference light intensity.

3. The method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination according to claim 2, characterized in that: In step S132, the ratio of the light intensity with the minimum duty cycle to the duty cycles of other light intensities is r.

4. The method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination according to claim 1, characterized in that: In step S2, when re-collecting the finger vein image, re-shoot k+1 groups of finger vein images with different light intensities, and calculate the grayscale value under any combination of duty cycle driving according to the principle of light intensity superposition. and Here are the steps: S21: According to the total light intensity E of each column of fingers j The pixel grayscale obtained by angiography is used to obtain the light intensity response equations of multiple K lamp combinations; S22: Solve the equations to find the minimum duty cycle of each lamp bead and the maximum grayscale value A of each column in the finger vein image. j and grayscale minimum value B j IRC offset value and S23: Calculate the two points A in each column j and B j Grayscale value under the duty cycle drive of any combination of K lamp beads and 5. The method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination according to claim 4, characterized in that: The total light intensity E of each row of fingers j It is K lamp beads with a light intensity of E jk The superposition expression is as follows: All pixels I(i,j) (i=1,2,…,M) in the jth column are illuminated by the uniform light intensity E of the kth lamp bead. jk When the contrast is in effect, the pixel grayscale obtained is expressed as follows: I(i,j,k)=f(E ijk +T i,j )j=1,2,…,N,k=1,2,…,K,i=1,2,…,M。 6. The method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination according to claim 5, characterized in that: In step S23, calculate the two points A in each column j and B j Grayscale value under the duty cycle drive of any combination of K lamp beads and Here are the steps: S231: Grayscale value and The expression is as follows: Two Points A j and B j The grayscale difference is expressed as follows: According to j = 1, 2, ..., N, and combined with the overall contrast of the finger vein image, the sum of multiple column contrasts C and The expression is as follows: in, 7. The method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination according to claim 6, characterized in that: In step S231, the overall contrast of the finger vein image is the sum C of the contrasts of multiple columns, which is expressed as follows:

8. The method for controlling the light intensity of a lamp bead to maximize the contrast of finger veins under non-uniform illumination according to claim 1, characterized in that: In step S3: Genetic algorithm is used to find the K LED light intensity combinations that reach the maximum value. The steps are as follows: According to the light intensity of K lamp beads in each column and the IRC offset of points A and B in each column of the finger image and The genetic algorithm is used to calculate the best combination and obtain the pk when the column grayscale contrast and value are the maximum; The actual circuit is driven to complete the finger vein image when the sum of the grayscale contrast of all pixel columns reaches the maximum value. k .

Images