CN107092879B - A method for monitoring fingerprint identification technology using near-infrared absorption - Google Patents

Abstract

The invention discloses a method for monitoring fingerprint identification technology by utilizing near-infrared absorption. Based on the infrared absorption characteristics, the best function matching of the data is obtained by measuring the optical absorbance of the trapping cross section in the wavelength range for curve fitting and least squares analysis, and the coherent characteristics such as extrema are obtained by using the optimal expression. Mathematical analysis. By comparing the database components, the verification of the identity of the living body is realized. Due to the thin vascular network in the superficial dermis and the physiological structural characteristics of large blood content and the absorption characteristics of hemoglobin and cytochrome in a specific near-infrared region, it cannot be imitated by other materials, and has anti-counterfeiting and non-replicability. At the same time, with the maturity of near-infrared spectroscopy technology, Fourier transform infrared spectrometers, raster scanners and other instruments can meet the relevant technical requirements. While ensuring efficiency, the method will greatly enhance the accuracy of identity recognition, strengthen information security, realize enhanced information security, and realize double detection of information.

Description

A method for monitoring fingerprint identification technology using near-infrared absorption

technical field

The invention relates to a method for monitoring fingerprint identification technology by utilizing near-infrared absorption, and belongs to the technical field of information security.

Background technique

In a highly information-based modern society, with the rapid development of communications, networks, and financial technologies, information security has shown unprecedented importance. In daily life and many occasions such as judicial, financial, security inspection, e-commerce, etc., more reliable, stable, and difficult to forge identification technology is required. In recent years, with the development of science and technology and computers, some biometric identification technologies such as face recognition, fingerprint recognition, and iris recognition have been popularized. Among them, fingerprint recognition has become the most widely used for its lifetime invariance, uniqueness and convenience. Personal identity authentication method. However, at this stage, fingerprint identification technology still has huge security risks. It cannot identify the ontology identity of the features to be extracted, lacks real-time monitoring of the transformation process of information and images, and the feature of the exposed body surface of fingerprints, which makes the information easy to be copied and stolen.

In 1977, Kaiser and Jobsis first reported the absorption characteristics of hemoglobin and cytochrome in a specific near-infrared region, and found that oxyhemoglobin and deoxyhemoglobin had two absorption peaks at 760 and 850 nm, and recorded the infrared spectrogram data through experimental measurement. . The present invention takes this as the technical background, uses curve fitting analysis to extract features, and monitors the fingerprint identification technology by verifying the identity of the living body.

SUMMARY OF THE INVENTION

In view of the deficiencies of the background technology, the present invention provides a method for monitoring fingerprint identification technology using near-infrared absorption, so as to improve the anti-counterfeiting property and increase the identification accuracy. The main steps of the technical solution are as follows:

In view of the deficiencies of the background technology, the present invention provides a method for monitoring fingerprint identification technology using near-infrared absorption, so as to improve the anti-counterfeiting property and increase the identification accuracy. The main steps of the technical solution are as follows:

In view of the deficiencies of the background technology, the present invention provides a method for monitoring the fingerprint identification technology by using near-infrared absorption, so as to improve the anti-counterfeiting property and increase the identification accuracy. The main steps of the technical solution are as follows:

Step 1) parameter initialization;

Step 2) Estimate the effective area, remove the edge of the image and the part with little change in gray level, and divide the fingerprint image into n×n (8≤n≤20) squares, where n represents the number of divisions; calculate each square The gray mean value and variance of the interval, if the two meet the conditions, the current square is defined as an effective square; connect all the effective squares and perform post-processing to obtain the effective area of the fingerprint, the record length l, l represents the length of the shorter side of the area. At the same time, the information of the effective area of the fingerprint is transmitted to the optical fiber probe, and the infrared transmitter automatically adjusts the emission angle to ensure that the emitted light enters the effective area;

step 3)

(3-1) Select the wavelength range [400, 1000] nm, control the light penetration thickness to reach the distance from the fingerprint contact surface within b*(0.3±0.05mm) to ensure that the supremum enters the dermis layer, use b to mark is the actual penetration thickness of light; (3-2) The incident light passes through the dermis medium, then

S=db·l

-dI _x =k·I _x

S represents the medium cross section, db represents the differentiation of b, -dI _x represents the absorbed light intensity, where I _x represents the light intensity radiated on the medium cross section S, k represents the probability that the light quantum is captured when it collides with the substance molecule; a Since any molecule has a trapping cross-section for light quantum, the total trapping area, that is, the effective area, is a·N, where N is the number of molecules in the medium cross-section S, so k=a·N/S

N=N _A ·c·10 ^-3 ·S·db

N _A is Avogadro's constant, c is the concentration of the substance, the unit is mol/L, so

-dIx _{=k·Ix =} (a·NA·c· ^10-3 · _S · _Ix /S)·db=( _a ·NA·c· _Ix /1000)·db

Take points on both sides

Have

ln(I ₀ /I)= _a ·NA·c·b/1000

Among them, I ₀ , I represent the intensity of incident light and the intensity of outgoing light, respectively;

Take the logarithm to base lg on both sides

lg(I ₀ /I)=a·NA·c·b/(2.303×10 ⁻³ )=2.64×10 ²⁰ _a ·c·b

The absorption K=lg(I ₀ /I), 2.64× ^{10 20} a is the molar absorption coefficient ε, then there is

K=εbc

(3-3) In the selected visible light and near-infrared wavelengths [400, 1000] nm, measure the discrete value of absorbance K _i corresponding to λ _i (i=1, 2, . . . , 600), where the subscript i means that the interval is evenly divided into 600 equal parts, λ _i means the wavelength value corresponding to the subscript i in the wavelength interval, and different wavelengths are selected for measurement to obtain a set of discrete data;

Step 4)

(4-1) Use curve fitting to compare the functional relationship between the discrete values of absorbance K _i and λ _i . The computer analyzes by the least squares method, and uses the sum of the squares of the minimized errors to find the best function matching of the data, set the function is f, that is

K ₁ =f(λ ₁ ,λ ₂ ,...,λ _T )

K ₂ =f(λ ₁ ,λ ₂ ,...,λ _T )

K ₃ =f(λ ₁ ,λ ₂ ,...,λ _T )

K ₄ =f(λ ₁ ,λ ₂ ,...,λ _T )

...

K _T =f(λ ₁ ,λ ₂ ,...,λ _T )

Among them, T represents the number of measurement groups, and T is not less than 100;

If λ _j is the true value, the true value y _j is obtained from the above-mentioned known function, and if its measured value is y _j ^* , the corresponding error is σ _j =y _j -y _j ^* (j=1, 2, .. ., n), using the least squares method

Obtain the sum of squares of the minimum error, where P _j represents the weight factor of each measurement value, and use the sum of squares of the minimum error to fit the optimal function expression f(λ);

(4-2) According to the expression, find the maximum value point and the coherence characteristic and perform the maximum likelihood estimation mathematical analysis, and determine whether it is blood in the finger by comparing the database components.

beneficial effect

(1) Using the thin vascular network in the superficial dermis of the finger and the physiological structure with a large blood content, by identifying the existence of the vascular network, the living body identity of the finger is effectively authenticated.

(2) Accurately exclude the identity of the prosthesis, such as fingerprint photos, fingerprint sleeves, fingerprint jammers, etc. currently on the market. Because the thickness level of the measurement area is 0.1mm, and the upper and lower bounds are specified, the internal characteristics of human skin cannot be imitated by other materials.

(3) At present, the near-infrared spectroscopy measurement technology is becoming more and more mature. Fourier transform infrared spectrometers, raster scanners, etc. can be imaged. The measurement of absorbance can be realized by computer, combined with fingerprint recognition and imaging, to effectively realize double verification of identity.

(4) It has a wide range of applications and promising prospects, and is suitable for daily life and many fields such as justice, finance, security inspection, and e-commerce.

Description of drawings

Fig. 1 is the transmission schematic diagram of light;

Figure 2 is the theoretical function of wavelength and absorbance;

Fig. 3 is the modeling flow chart of fingerprint identification by near-infrared absorption monitoring.

Detailed ways

The technical solutions of the present invention will be described in detail below with reference to the accompanying drawings.

A method for monitoring fingerprint identification technology using near-infrared absorption, comprising the following steps:

Step 1) parameter initialization;

step 2)

Perform effective area estimation, remove image edges and parts with little change in gray level, and divide the fingerprint image into n×n (8≤n≤20) squares, where n represents the number of divisions; calculate the gray level of each square interval. Degree mean and variance, if the two meet the conditions, the current square is defined as a valid square; connect all valid squares and perform post-processing to obtain the fingerprint valid area, record length l, l represents the length of the shorter side of the area; at the same time , transmit the information of the effective area of the fingerprint to the fiber probe, and the infrared transmitter automatically adjusts the emission angle to ensure that the emitted light enters the effective area;

Step 3) (3-1) Select the wavelength range [400,1000] nm, and control the light penetration thickness to reach the distance from the fingerprint contact surface within b*(0.3±0.05mm) to ensure that the supremum enters the dermis layer, Use b as the actual penetration thickness of light;

(3-2) The incident light passes through the dermis medium, then

S=db·l

-dI _x =k·I _x

S represents the medium cross section, db represents the differentiation of b, -dI _x represents the absorbed light intensity, where I _x represents the light intensity radiated on the medium cross section S, k represents the probability that the light quantum is captured when it collides with the substance molecule; a Since any molecule has a trapping cross-section for light quantum, the total trapping area, that is, the effective area, is a·N, where N is the number of molecules in the medium cross-section S, so k=a·N/S

N=N _A ·c·10 ^-3 ·S·db

N _A is Avogadro's constant, c is the concentration of the substance, the unit is mol/L, so

-dIx _{=k·Ix =} (a·NA·c· ^10-3 · _S · _Ix /S)·db=( _a ·NA·c· _Ix /1000)·db

Take points on both sides

Have

ln(I ₀ /I)= _a ·NA·c·b/1000

Among them, I ₀ , I represent the intensity of incident light and the intensity of outgoing light, respectively;

Take the logarithm to base lg on both sides

lg(I ₀ /I)=a·NA·c·b/(2.303×10 ⁻³ )=2.64×10 ²⁰ _a ·c·b

The absorption K=lg(I ₀ /I), 2.64× ^{10 20} a is the molar absorption coefficient ε, then there is

K=εbc

(3-3) In the selected visible light and near-infrared wavelengths [400, 1000] nm, measure the discrete value of absorbance K _i corresponding to λ _i (i=1, 2, . . . , 600), where the subscript i means that the interval is evenly divided into 600 equal parts, λ _i means the wavelength value corresponding to the subscript i in the wavelength interval, and different wavelengths are selected for measurement to obtain a set of discrete data;

Step 4)

(4-1) Use curve fitting to compare the functional relationship between the discrete values of absorbance K _i and λ _i . The computer analyzes by the least squares method, and uses the sum of the squares of the minimized errors to find the best function matching of the data, set the function is f, that is

K ₁ =f(λ ₁ ,λ ₂ ,...,λ _T )

K ₂ =f(λ ₁ ,λ ₂ ,...,λ _T )

K ₃ =f(λ ₁ ,λ ₂ ,...,λ _T )

K ₄ =f(λ ₁ ,λ ₂ ,...,λ _T )

...

K _T =f(λ ₁ ,λ ₂ ,...,λ _T )

Among them, T represents the number of measurement groups, and T is not less than 100;

If λ _j is the true value, the true value y _j is obtained from the above-mentioned known function, and if its measured value is y _j ^* , the corresponding error is σ _j =y _j -y _j ^* (j=1, 2, .. ., n), using the least squares method

Obtain the sum of squares of the minimum error, where P _j represents the weight factor of each measurement value, and use the sum of squares of the minimum error to fit the optimal function expression f(λ);

(4-2) According to the expression, find the maximum value point and the coherence characteristic and perform the maximum likelihood estimation mathematical analysis, and determine whether it is blood in the finger by comparing the database components.

Claims (1)

1. utilize the method for near-infrared absorption monitoring fingerprint identification technology, is characterized in that, comprises the steps: Step 1) parameter initialization; Step 2) Estimate the effective area, remove the edge of the image and the part with little change in gray level, and divide the fingerprint image into n×n, 8≤n≤20 squares, where n represents the number of divisions; calculate the interval of each square If the two meet the conditions, the current square is defined as the effective square; connect all the effective squares and perform post-processing to obtain the fingerprint effective area, the record length l, l represents the length of the shorter side of the area ; At the same time, the information of the effective area of the fingerprint is transmitted to the optical fiber probe, and the infrared transmitter automatically adjusts the emission angle to ensure that the emitted light enters the effective area; step 3) (3-1) Select the wavelength range [400, 1000] nm, and control the light penetration thickness to reach the distance b* from the fingerprint contact surface, that is, the value is within 0.3±0.05mm, to ensure that the supremum enters the dermis layer; use b Recorded as the actual penetration thickness of light; (3-2) The incident light passes through the dermis medium, then S=db·l -dI _x =k·I _x S represents the medium cross section, db represents the differentiation of b, -dI _x represents the absorbed light intensity, where I _x represents the light intensity radiated on the medium cross section S, k represents the probability that the light quantum is captured when it collides with the substance molecule; a Since any molecule has a trapping cross-section for photons, the total trapping area, that is, the effective area, is a·N, where N is the number of molecules in the medium cross-section S, so k=a·N/S N=N _A ·c·10 ^-3 ·S·db N _A is Avogadro's constant, c is the concentration of the substance, the unit is mol/L, so -dIx _{=k·Ix =} (a·NA·c· ^10-3 · _S · _Ix /S)·db=( _a ·NA·c· _Ix /1000)·db Take points on both sides

Have ln(I ₀ /I)= _a ·NA·c·b/1000 Among them, I ₀ , I represent the intensity of incident light and the intensity of outgoing light, respectively; Take the logarithm to base lg on both sides lg(I ₀ /I)=a·NA·c·b/(2.303×10 ⁻³ )=2.64×10 ²⁰ _a ·c·b The absorption K=lg(I ₀ /I), 2.64× ^{10 20} a is the molar absorption coefficient ε, then there is K=εbc (3-3) In the wavelength range of [400, 1000] nm of selected visible light and near-infrared light, measure the discrete value of absorbance K _i corresponding to λ _i , i=1, 2,..., 600, where the subscript i Indicates that the interval is evenly divided into 600 equal parts, λ _i represents the wavelength value corresponding to the subscript i in the wavelength interval, and different wavelengths are selected for measurement to obtain a set of discrete data; Step 4: (4-1) Use curve fitting to compare the functional relationship between the discrete values of absorbance K _i and λ _i . The computer analyzes by the least squares method, and uses the sum of the squares of the minimized errors to find the best function matching of the data, set the function is f, that is K ₁ =f(λ ₁ ,λ ₂ ,...,λ _T ) K ₂ =f(λ ₁ ,λ ₂ ,...,λ _T ) K ₃ =f(λ ₁ ,λ ₂ ,...,λ _T ) K ₄ =f(λ ₁ ,λ ₂ ,...,λ _T ) ... K _T =f(λ ₁ ,λ ₂ ,...,λ _T ) Among them, T represents the number of measurement groups, and T is not less than 100; If λ _j is a true value, the true value y _j is obtained from the above-mentioned known function, and if its measured value is y _j ^* , the corresponding error is σ _j =y _j -y _j ^* j = 1, 2, ... , n, using the least squares method

Obtain the sum of squares of the minimum error, where P _j represents the weight factor of each measurement value, and use the sum of squares of the minimum error to fit the optimal function expression f(λ); (4-2) According to the expression, find the maximum value point and the coherence characteristic and perform the maximum likelihood estimation mathematical analysis, and determine whether it is blood in the finger by comparing the database components.

Images