A method for maximizing local contrast of finger vein images under uniform illumination

Abstract

The present invention discloses a method for maximizing the local contrast of a finger vein image under uniform illumination, and the steps are as follows: designing a function model of a vein point illumination response curve with unknown parameters under uniform near-infrared illumination; using a uniform illumination control system to control the emission of N kinds of near-infrared light of illumination intensity to obtain N finger vein images; obtaining the following relationship between the function model of the illumination response curve of a non-vein point and the function model of the illumination response curve of a vein point: selecting M illumination intensity values ​​to obtain 2M intersections of the curve; then solving the unknown parameters; obtaining the function model of the final vein point illumination response curve f(E); solving the optimal illumination intensity point E corresponding to the maximum local contrast of the finger vein. m , and obtain the current finger vein local contrast maximized image.

Description

Technical Field

[0001] The present invention relates to the technical field of finger vein image processing, and in particular to a method for maximizing the local contrast of a finger vein image under uniform illumination. Background Art

[0002] In the prior art, near-infrared LED lamp arrays are used to collect finger vein images. The light intensity distribution is uneven when the light is irradiated on the back of the finger through the lamp array. After the near-infrared light is absorbed, diffused and reflected by the bones, muscles, fat, body fluids and other physiological tissues inside the finger, it reaches the superficial subcutaneous part of the inner side of the finger for venography. That is, the light intensity is stronger directly below the LED lamp bead, and the light gradually weakens as the distance from the LED lamp bead increases. When controlling the light intensity, a larger redundancy range of light intensity is required, which is difficult to adjust and easily causes overexposure or underexposure.

[0003] Furthermore, even if overexposure and underexposure are avoided during use, due to the low contrast of the vein patterns in the finger vein image, the subsequent local grayscale increase, local histogram equalization and other methods required to enhance the contrast of the finger vein patterns during image data processing require large amounts of calculations, which is not conducive to practical applications. Existing methods rarely study how to maximize the grayscale contrast of the vein patterns during the acquisition process, because the illumination is uneven and the thickness of the finger is unevenly distributed. Under such conditions, it is difficult to model the relationship between the illumination and imaging grayscale of the finger veins and to maximize the local contrast. Summary of the invention

[0004] In view of this, the present invention provides a method for maximizing the local contrast of finger vein images under uniform illumination, designs a uniform illumination control system, and solves the optimal illumination intensity point E corresponding to the maximum local contrast of finger veins. m , to capture the best finger vein image, thereby obtaining an image with maximized local contrast of the finger vein.

[0005] The technical solution is as follows:

[0006] A method for maximizing the local contrast of a finger vein image under uniform illumination, characterized by following the steps below:

[0007] S1: Design a function model of the illumination response curve f(E) of the vein point with unknown parameters under uniform near-infrared illumination;

[0008] S2: Using a uniform illumination control system to emit near-infrared light of N different illumination intensities, and obtaining N finger vein images;

[0009] S3: According to the N finger vein images obtained in step S2, grayscale finger vein images are obtained after grayscale processing, and according to the local grayscale values ​​in the grayscale finger vein images, the non-vein point illumination response curve F(E) and the vein point illumination response curve f(E) are fitted; and it is concluded that the function model of the non-vein point illumination response curve F(E) and the function model of the vein point illumination response curve f(E) have the following relationship:

[0010] F(E) = f(E+t);

[0011] t is the light intensity shift value between the light response curve of the non-vein point and the light response curve of the vein point, which is an unknown parameter;

[0012] E is the light intensity;

[0013] According to the theory proposed in the paper Paul E. Debevec, Jitendra Malik. Recovering high dynamic range radiance maps from photographs [P]. Computer graphics and interactive techniques, 1997, the illumination response curve of each pixel is similar, which is a translation of the camera response curve. Here, the illumination response curve f(E) of the vein point is taken as the benchmark, and the illumination response curve F(E) of the non-vein point is considered to differ from it by a translation t, that is, F(E) = f(E+t).

[0014] S3: According to the number of all unknown parameters in the function model for solving the venous point illumination response curve f(E), M illumination intensity values ​​are selected to obtain 2M intersection points corresponding to the M illumination intensity values ​​on the non-venous point illumination response curve F(E) and the venous point illumination response curve f(E);

[0015] S4: Substituting the values ​​of the 2M intersection points into the non-venous point illumination response curve F(E) and the vein point illumination response curve f(E), solving the unknown parameters in the function model of the vein point illumination response curve f(E); obtaining the final function model of the vein point illumination response curve f(E);

[0016] S5: Define the local contrast of finger veins as the grayscale difference between the points on the local non-vein and the points on the vein, and find the optimal illumination intensity point E corresponding to the maximum local contrast of finger veins. m ;

[0017] S6: The uniform illumination control system controls the optimal illumination intensity point E calculated in step S5. m After emitting near-infrared light and transmitting it through the finger, an image with the maximum local contrast of the current finger vein is obtained.

[0018] Through the above design, a uniform illumination control system is used to emit uniform illumination, which avoids uneven illumination and makes it difficult to model the contrast and obtain the maximum value. With uniform illumination, the camera illumination response model can be used for modeling, and the illumination intensity that maximizes the contrast can be further obtained, thereby obtaining a more accurate finger illumination response curve. In the process of finding the optimal illumination intensity, the illumination response curve is modeled, and the light intensity is adjusted using the illumination response curve. Through multiple illumination and image acquisition, a finger vein image with the maximum local vein contrast in theoretical calculation is obtained.

[0019] Among them, the non-vein point illumination response curve F(E) is the fitting curve of the local grayscale maximum point;

[0020] The vein point illumination response curve f(E) is the fitting curve of the local grayscale minimum point.

[0021] Furthermore, the function model of the finger vein illumination response curve f(E) with unknown parameters is:

[0022] f(E)=2 / π*arctan(a*(b*E+1 / 2) c );

[0023] Where E is the light intensity; a, b and c are all unknown parameters that fit discrete points to a fixed value range.

[0024] Furthermore, in the function model of the finger vein illumination response curve f(E), the strongest illumination intensity emitted by the normalized light source is set to 1, and the illumination intensity when the light source is not emitting light is set to 0, so 0≤E≤1;

[0025] f(E) is the grayscale value of a pixel (x, y) under light intensity E, which is normalized from [0,255] to [0,1];

[0026] 2 / π is a fixed coefficient. Since the value range of the arctan function is (-π / 2,π / 2), it is multiplied by 2 / π and normalized to [0,1] after rounding both ends. That is, if it is less than 1 / 255, it is 0, and if it is greater than 254 / 255, it is 1.

[0027] Furthermore, in step S5, the optimal illumination intensity point E corresponding to the maximum local contrast of the finger vein is solved. m The steps are:

[0028] The intersection of the non-vein point illumination response curve F(E) and the vein point illumination response curve f(E) is the optimal illumination intensity point E corresponding to the maximum contrast of the finger vein. m , that is: F′(Em )-f′(E m )=0.

[0029] Using the above scheme, the contrast of local veins is defined as the grayscale difference between points on the local non-venous point and the points on the vein. Under a certain light intensity, the quantized contrast value is F(E)-f(E), and its maximization can be derived.

[0030] Furthermore, the uniform illumination control system includes a filter and an LED light guide plate arranged opposite to each other, the space between the filter and the LED light guide plate is a finger placement space, the side of the LED light guide plate away from the filter is a shooting surface, a camera is fixed on the shooting surface side of the LED light guide plate, and the camera is aimed at the finger placement space; the camera is connected to the acquisition control end of the microcontroller; the driving end of the microcontroller is connected to the LED light guide plate light intensity driving circuit, and the driving output end of the LED light guide plate light intensity driving circuit is connected to the LED light guide plate.

[0031] With the above solution, the dots of the light source on the side of the light guide plate close to the LED are small, while the dots of the light source on the side far from the LED are large. When the light source enters the light guide plate and hits the reflection point, the light on the part of the light guide plate closer to the light source is stronger, and the light on the part of the light guide plate far from the light source is weaker. The two are mutually reinforcing, and optical simulation software can be used to design a near-infrared light with uniform intensity throughout the light guide plate.

[0032] The LED light guide plate is placed directly against the back of the finger, and a uniform illumination control system forms a control closed loop from emitting near-infrared light to capturing the finger image, which is conducive to real-time adjustment of the illumination intensity according to the acquired finger image.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] This patent models the illumination response curve of the camera under uniform illumination conditions, uses the illumination response curve to adjust the light intensity, and obtains a finger vein image that maximizes the local vein contrast in theoretical calculations through multiple illumination and image acquisition. Compared with the binary method in the prior art, the number of light intensity updates is significantly reduced; compared with the Newton method, the amount of calculation is significantly reduced. It is a method that takes into account the advantages of both the binary method and the Newton method and avoids the disadvantages of both, and has obvious advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A flowchart of maximizing the local contrast of a finger vein image according to the present invention;

[0036] Figure 2 This is a control block diagram of the uniform illumination control system of the present invention;

[0037] Figure 3 It is a brief schematic diagram of the uniform illumination control system of the present invention;

[0038] Figure 4 A schematic diagram of the illumination response curves of the vein point and the non-venous point in the design model state of the present invention;

[0039] Figure 5 A schematic diagram of the derivative function of the venous point curve and the non-venous point curve of the present invention;

[0040] Figure 6 The local contrast maximization point E m Result graph;

[0041] Figure 7 It is a schematic diagram of the illumination response curve of the vein point and non-venous point under special conditions;

[0042] Figure 8 It is a schematic diagram of the illumination response curves of the vein point and the non-venous point under the general solution;

[0043] Fig. 9 Schematic diagram of illumination curves of multiple vein points and non-venous points in the general solution;

[0044] Fig.10 It is a schematic diagram of illumination curves of multiple vein points and illumination curves of non-venous points under the design model state of the present invention. DETAILED DESCRIPTION

[0045] The present invention is further described below in conjunction with embodiments and drawings.

[0046] A method based on maximizing the local contrast of finger vein images under uniform illumination, combined with Figure 1 It can be seen that the following steps are followed:

[0047] S1: Design a function model of the illumination response curve f(E) of the vein point with unknown parameters under uniform near-infrared illumination;

[0048] In this embodiment, the function model of the unknown parameter venous illumination response curve f(E) is:

[0049] f(E)=2 / π*arctan(a*(b*E+1 / 2) c );

[0050] Where E is the light intensity; a, b and c are all unknown parameters that fit discrete points to a fixed value range.

[0051] S2: Using a uniform illumination control system to emit N kinds of infrared light with similar illumination intensities, N finger vein images are obtained;

[0052] In this embodiment, see Figure 2 and 3 It can be seen that the uniform illumination control system includes a filter 3 and an LED light guide plate 1 which are arranged opposite to each other, the space between the filter 3 and the LED light guide plate 1 is a finger placement space, the side of the LED light guide plate 1 away from the filter 3 is a shooting surface, a camera 4 is fixed on the shooting surface side of the LED light guide plate 1, and the camera 4 is aimed at the finger placement space; the camera 4 is connected to the acquisition control end of the microcontroller 5; the driving end of the microcontroller 5 is connected to the LED light guide plate light intensity driving circuit 6, and the driving output end of the LED light guide plate light intensity driving circuit 6 is connected to the LED light guide plate 1.

[0053] In this embodiment, the USB interface of the microcontroller 5 is used as the acquisition control terminal. S3: According to the N finger vein images obtained in step S2, see Figure 4 , the non-venous point illumination response curve F(E) and the venous point illumination response curve f(E) are fitted; and the function model of the non-venous point illumination response curve F(E) and the function model of the venous point illumination response curve f(E) are obtained, and the following relationship exists between them:

[0054] F(E) = f(E+t);

[0055] t is the light intensity shift value between the light response curve of the non-vein point and the light response curve of the vein point, which is an unknown parameter;

[0056] E is the light intensity;

[0057] In the function model of the venous illumination response curve f(E), the strongest illumination intensity that the normalized light source can emit is set to 1, and the illumination intensity when the light source is not emitting light is set to 0, so 0≤E≤1;

[0058] f(E) is the grayscale value of a pixel (x, y) under light intensity E, which is normalized from [0,255] to [0,1];

[0059] 2 / π is a fixed coefficient. Since the value range of the arctan function is (-π / 2,π / 2), it is multiplied by 2 / π and normalized to [0,1] after rounding both ends.

[0060] S3: According to the number of all unknown parameters in the function model for solving the venous point illumination response curve f(E), M illumination intensity values ​​are selected to obtain 2M intersection points corresponding to the M illumination intensity values ​​on the non-venous point illumination response curve F(E) and the venous point illumination response curve f(E);

[0061] In this embodiment, M=2.

[0062] See also Figure 4 , the horizontal axis E of the response curve is the light intensity, and the vertical axis G is the grayscale value; the two light intensity values ​​are: low light intensity E A and high light intensity E B ; The four points selected are: G AV :Low light intensity E in the vein area A The next point, G AU : Low light intensity E in non-venous area A The next point, G BV : High light intensity E in the vein area B The next point, G BU : High light intensity E in non-venous area B The next point

[0063] S4: Substituting the values ​​of the 2M intersection points into the non-venous point illumination response curve F(E), solving the unknown parameters in the function model of the vein point illumination response curve f(E); obtaining the final function model of the vein point illumination response curve f(E);

[0064] That is, according to the existing function and the four points obtained under two different light intensities, that is, G AV , G AU , G BV , G BU , the four parameters a, b, c and t can be solved.

[0065] In this embodiment, according to experimental data, the drawn camera response curve itself is fitted by a series of discrete points, and there is a certain error between the discrete points and the real data. Therefore, in order to reduce the impact of this error on the subsequent model solution parameters, it is necessary to ensure that there is a certain distance difference between the two points obtained under the same illumination intensity. From multiple experiments, the value interval of two different illumination intensities is set to [0.1 0.6], and the difference between the two illumination intensities is [0.05 0.5], which can achieve better results.

[0066] In this embodiment, the function model of the finger vein illumination response curve f(E) obtained by solving is:

[0067] f(E)=2 / π*arctan(0.001*π*(40*E+1 / 2) 3 );

[0068] Where 0≤E≤1.

[0069] S5: Define the local contrast of finger veins as the grayscale difference between points on non-veins and points on veins, and find the optimal illumination intensity point E corresponding to the maximum local contrast of finger veins. m ;

[0070] In step S5, the optimal light intensity point E corresponding to the maximization of the local contrast of finger veins is solved m The steps are as follows:

[0071] The intersection point of the curves obtained by differentiating the non-vein point light response curve F(E) and the vein point light response curve f(E) respectively is the optimal light intensity point E corresponding to the maximization of the local contrast of finger veins m , that is: F′(E m ) - f′(E m ) = 0.

[0072] Combined with Figure 5 , it is known that when 0 < E < Em, f’ < F’, and when E m < E < E B , f’ > F’, so there is f’ - F’ = 0 at E m .

[0073] S6: The uniform light control system controls the near-infrared light to be emitted at the optimal light intensity point E calculated according to step S5 m , and after passing through the finger, the current image with the maximized local contrast of finger veins is obtained.

[0074] Combined with Figure 6 , it is known that there is a maximum difference between f(E m ) and F(E m ) at E m ), that is, the maximization of the local contrast of finger veins proposed in this patent is ensured at the optimal light intensity.

[0075] In this patent, according to the function analytical formulas of F(E) and f(E), it is known that Figure 6 , when E A = 0.1, G AV = 0.1775, G AU = 0.4532; when E B = 0.345, G BV = 0.931, G BU = 0.9533; when Em = 0.115, F(E) = 0.5372, f(E) = 0.2514, F(E) - f(E) = 0.2858.

[0076] In the present invention, it should be supplemented that: in this patent, the difference between f(E) and F(E) is greater than the gray difference corresponding to a certain pixel point on any other two light response curves of the vein region and the non-vein region.

[0077] The maximum value of the local contrast of finger veins exists. First, the following special cases are excluded;

[0078] In the same range of light intensity E, the light response curves of non-vein points and vein points in special cases are as follows: Figure 7 shown.

[0079] exist Figure 7 Although the situation in the above example is consistent with mathematical logic, it is not consistent with the actual situation. Because in reality, there will be no non-venous area being illuminated. Figure 7 At point A in the figure (the grayscale value reaches 255), the vein area still has no near-infrared light passing through. Figure 7 The situation of point B in (the gray value is still 0).

[0080] Proof: In the present invention, the difference between f(E) and F(E) is greater than or equal to the grayscale difference corresponding to a pixel point on any other two illumination response curves of the vein region and the non-venous region.

[0081] As a general solution, the illumination response curves of non-venous points and vein points are Figure 8 shown.

[0082] Proof 1: The illumination response function / curve is monotonically increasing and is divided into underexposure region, normal working region and oversaturation region.

[0083] Order Figure 8 Among them, [0,E0] is working area 1, (E0,E1] is working area 2, (E1,E2] is working area 3, (E2,E3] is working area 4, and (E3,E] is working area 5.

[0084] In working area 1, both the non-vein point and the vein point are underexposed, and the grayscale difference corresponding to a certain pixel point on the illumination response curve between the vein point and the non-vein point is 0;

[0085] In working area 2, non-vein points are located in the normal area, and vein points are in the underexposed area;

[0086] In working area 3, both non-vein points and vein points are located in the normal working area;

[0087] In working area 4, the oversaturated area of ​​non-venous points overlaps with the normal working area of ​​the vein area;

[0088] In working area 5, both the non-vein point and the vein point belong to the over-exposed area, and the over-saturated area of ​​the non-vein point overlaps with the over-saturated area of ​​the vein point. As before, the difference between the two is 0.

[0089] In the working areas 1 and 5 described above, the gray-scale difference corresponding to a certain pixel point on the light response curves (F(E) and f(E)) of non-venous points and venous points is 0. Also, since the light response curves of non-venous points and venous points are both monotonically increasing, there must be a maximum difference in working areas 2, 3, and 4. And from Figure 8 it can be obtained that the difference between F(E) and f(E) in working area 2 is increasing, and the difference between F(E) and f(E) in working area 4 is decreasing. Therefore, the maximum value of the difference between F(E) and f(E) must be within the closed interval of working area 3.

[0090] Proof 2: Under the light conditions of working area 3, when the light response curves of other points (such as f1, F1) are located between the light response curve of non-venous points F (maximum gray-scale pixel) and the light response curve of venous points f (minimum gray-scale pixel) proposed in this model, the difference (F1 - f1) is less than or equal to F(E) - f(E). The proof is as follows:

[0091] Among them, the light response curve of venous points is f(E) = 2 / π * arctan(a * π * (b * E + 1 / 2) c ), and the light response curve of non-venous points is F(E) = 2 / π * arctan(a * π * (b * (E + t) + 1 / 2) c ).

[0092] Now it is known that the auxiliary venous point curve and the auxiliary non-venous point curve are always located between f(E) and F(E). Assume that there are any number of auxiliary venous point light curves and any number of auxiliary non-venous point light curves between f(E) and F(E).

[0093] Let one of the auxiliary venous point light response curves be:

[0094] f1(E) = f(E + d) = 2 / π * arctan(0.001 * π * (40 * (E + d) + 1 / 2) 3 );

[0095] Let one of the auxiliary non-venous point light response curves be:

[0096] F(E) = f(E + w) = 2 / π * arctan(0.001 * π * (40 * (E + w) + 1 / 2) 3 ).

[0097] From Fig. 9 it is obvious that when E is in working area 3, that is, E1 < E < E2, F’(E) > F1’(E) > f1’(E) > f’(E) > 0, that is Fig. 9 the drawn curves are all in a monotonically increasing state when E1 < E < E2. Take an E value in working area 3 and let E = EA , then F(E A )>F1(E A )>f1(E A )>f(E A )>0,F(E A )-f(E A )>F1(E A )-f1(E A ), the local contrast of the finger vein obtained by the model of the present invention (the difference between the non-vein point curve F(E) and the vein point curve f(E)) is the maximum difference between any two non-vein point curves and the vein point curve.

[0098] Proof 3:

[0099] Under the same illumination intensity E, when E is located in the working area 3 and E∈[0.1,0.6], compared with any other two non-vein point and vein point illumination response curves, the difference between the non-vein point illumination response curve F(E) and the vein point illumination response curve f(E) obtained by the model proposed in this patent is greater than the grayscale difference corresponding to a certain pixel point on any other two non-vein point and vein point illumination response curves. The proof is as follows:

[0100] In the model designed by the present invention, it is taken as a special solution. The illumination response curve of the vein point is:

[0101] f(E)=2 / π*arctan(0.001*π*(40*E+1 / 2) 3 );

[0102] The light response curve of the non-venous point is F(E) = f(E+2) = 2 / π*arctan(0.001*π*(40*(E+2)+1 / 2) 3 );

[0103] like Fig.10 As shown, it can be seen that the f(E) and F(E) curves show a strictly monotonically increasing trend. To prove: Under the same light intensity E, when E∈[0.1,0.6], compared with the illumination response curves of multiple other veins and non-vein points, the difference between f(E) and F(E) in the present invention is greater than the grayscale difference corresponding to a pixel point on any other two illumination response curves of vein areas and non-vein areas. The proof is as follows:

[0104] It is known that the auxiliary vein point curve and the auxiliary non-vein point curve are always located between f(E) and F(E). Assume that there are n auxiliary vein point illumination curves and n auxiliary non-vein point illumination curves between f(E) and F(E).

[0105] Assume that the illumination response curve of one of the vein points is:

[0106] f1(E)=f(E+1 / 2)=2 / π*arctan(0.001*π*(40*(E+1 / 2)+1 / 2) 3 );

[0107] One of the non-vein point illumination response curves is:

[0108] F1(E)=f(E+1)=2 / π*arctan(0.001*π*(40*(E+1)+1 / 2) 3 );

[0109] When E∈[0,1], the illumination response curves of f(E), F(E), f1(E) and F1(E) are all monotonically increasing;

[0110] When E∈[0,0.1], the illumination response curves of f(E), F(E), f1(E) and F1(E) are in the underexposure range;

[0111] When E∈[0.6,1], the light response curves of f(E), F(E), f1(E) and F1(E) are in the overexposure range. Therefore, when E∈[0.1,0.6], F(E)>F1(E)>f1(E)>f(E)>0;

[0112] When E=0.1, let E=E A , F(E A )=G AU ,F1(E A )=G AU1 ,f1(E A )=G AV1 ,f(E A )=G AV , so when E=0.1, G AU -G AV >G AU1 -G AV1 .

[0113] Then, by the same token, when E∈[0.1,0.6], G AU -G AV >G AUi -G AVi , i∈[1,n].

[0114] That is, the local contrast of the finger vein obtained by the model of the present invention (the difference between the non-vein point curve F(E) and the vein point curve f(E)) is the maximum difference between any two non-vein point curves and the vein point curve.

[0115] Finally, it should be noted that the above description is only a preferred embodiment of the present invention. Under the guidance of the present invention, ordinary technicians in this field can make various similar expressions without violating the purpose and claims of the present invention, and such changes all fall within the scope of protection of the present invention.

Claims

1. A method for maximizing the local contrast of a finger vein image under uniform illumination, characterized in that: Follow these steps: S1: Design a function model of the illumination response curve f(E) of the vein point with unknown parameters under uniform near-infrared illumination; S2: Using a uniform illumination control system to emit near-infrared light of N different illumination intensities, and obtaining N finger vein images; S3: According to the N finger vein images obtained in step S2, the non-vein point illumination response curve F(E) and the vein point illumination response curve f(E) are fitted; and it is concluded that the function model of the non-vein point illumination response curve F(E) and the function model of the vein point illumination response curve f(E) have the following relationship: F(E) = f(E+t); t is the light intensity shift value between the light response curve of the non-vein point and the light response curve of the vein point, which is an unknown parameter; E is the light intensity; S3: According to the number of all unknown parameters in the function model for solving the venous point illumination response curve f(E), M illumination intensity values ​​are selected to obtain 2M intersection points corresponding to the M illumination intensity values ​​on the non-venous point illumination response curve F(E) and the venous point illumination response curve f(E); S4: Substituting the values ​​of the 2M intersection points into the non-venous point illumination response curve F(E) and the vein point illumination response curve f(E), solving the unknown parameters in the function model of the vein point illumination response curve f(E); obtaining the final function model of the vein point illumination response curve f(E); S5: Define the local contrast of finger veins as the grayscale difference between the points on the local non-vein and the points on the vein, and find the optimal illumination intensity point E corresponding to the maximum local contrast of finger veins. m ; S6: The uniform illumination control system controls the optimal illumination intensity point E calculated in step S5. m After emitting near-infrared light and transmitting it through the finger, an image with the maximum local contrast of the current finger vein is obtained; The function model of the unknown parameter venous illumination response curve f(E) is: f(E)=2 / π*arctan(a*(b*E+1 / 2) c ) Where E is the light intensity; a, b and c are all unknown parameters that fit discrete points to a fixed value range.

2. The method for maximizing local contrast of finger vein images under uniform illumination according to claim 1, characterized in that: In the function model of the finger vein illumination response curve f(E), the strongest illumination intensity emitted by the normalized light source is set to 1, and the illumination intensity when the light source is not emitting light is set to 0, so 0≤E≤1; f(E) is the grayscale value of a pixel (x, y) under light intensity E, which is normalized from [0,255] to [0,1]; 2 / π is a fixed coefficient. Since the value range of the arctan function is (-π / 2,π / 2), it is multiplied by 2 / π and normalized to [0,1] after rounding both ends.

3. The method for maximizing local contrast of finger vein images under uniform illumination according to claim 2, characterized in that: In step S5, the optimal illumination intensity point E corresponding to the maximum local contrast of the finger vein is solved. m The steps are: The intersection of the non-vein point illumination response curve F(E) and the vein point illumination response curve f(E) is the optimal illumination intensity point E corresponding to the maximum contrast of the finger vein. m , that is: F′(E m )-f′(E m )=0.

4. The method for maximizing local contrast of finger vein images under uniform illumination according to claim 1, characterized in that: The uniform illumination control system comprises a filter and an LED light guide plate arranged opposite to each other, the space between the filter and the LED light guide plate is a finger placement space, the side of the LED light guide plate away from the filter is a shooting surface, a camera is fixed on the shooting surface side of the LED light guide plate, and the camera is aimed at the finger placement space; the camera is connected to the capture control end of the microcontroller; the driving end of the microcontroller is connected to the LED light guide plate light intensity driving circuit, and the driving output end of the LED light guide plate light intensity driving circuit is connected to the LED light guide plate.

Images