CN101201893A - Iris recognition preprocessing method based on gray information - Google Patents

Abstract

The invention provides an iris image preprocessing method based on gray scale information, which firstly positions the rough circle center (x) of a pupil through operations such as binaryzation, mathematical morphology, gray scale projection and the like_o，y_o) Then at (x)_o，y_o) Searching for the sum-and-sum (x) with the gray value larger than T from the center to two sides on a plurality of nearby rows_o，y_o) The point with the shortest distance is taken as the boundary point of the pupil, the precise circle center and the radius of the pupil can be positioned through curve fitting, and thenAnd calculating the gray level first-order difference value of the row where the pixel points are searched, searching the maximum sum of the horizontal first-order difference values in the range where the iris outer boundary points possibly appear, taking the maximum sum as the iris outer boundary points, and performing curve fitting to obtain the accurate circle center and radius of the iris outer boundary. And finally, according to the normalized image, judging the quality of the iris image by calculating the point sharpness and the number of effective pixel points. By adopting the iris image preprocessing method, the time required by repeated iteration during positioning can be reduced, and the quality of the iris image can be judged quickly and accurately according to the extracted iris image.

Description

A preprocessing method for iris recognition based on grayscale information

technical field

The invention belongs to the technical field of image processing, and mainly relates to iris identification technology in biometric identification.

Background technique

In today's information age, how to accurately identify a person's identity and protect information security is a key social problem that must be solved. For this reason, biometric identification technology has quietly emerged, and has become a frontier research topic in the field of information security management in the world. Biometric identification technology refers to the use of the inherent physiological or behavioral characteristics of the human body for personal identification. Iris identification technology is a branch of biometric identification technology. It is the application of computer image processing technology and pattern recognition technology in the field of personal identification. Due to its high stability and high accuracy, it has become a popular development of biometric identification in recent years. direction. The iris automatic identification technology is widely used in banks, public security, airports, networks, etc., and has great economic value and practical significance. Now it has been used in border inspection, comprehensive authentication, bank withdrawal, information management and building security management, etc. It can also free people from the cumbersome memory of credit card numbers, bank account numbers, ID card numbers, and network login numbers. With the development of digital signal processing technology and image processing technology, iris identification system has become increasingly mature. For details, see: John G. Daugman, "How Iris Recognition Works," IEEETransaction on Circuits and Systems for Video Technology, Volume 14, Issue 1, pp.21-30, 2004 and: John G. Daugman, "High Confidence Recognition of Persons by Iris Patterns,” The Proceeding of IEEE 35 ^th International Carnahan Conference on Security Technology, pp.254-263, 2001.

In iris identification technology, iris image preprocessing is the key to the whole identification technology, which includes iris location and iris image quality evaluation. Iris location is the first step in iris recognition, and its execution time and accuracy will directly affect the speed and accuracy of the entire iris recognition system. In practice, since the iris area is often blocked by eyelids and eyelashes, the accuracy and effectiveness of the iris location algorithm needs to be further improved. How to quickly and accurately locate the iris in the low-quality iris image with eyelashes and eyelid occlusion, and describe its boundary or position with a mathematical model is the main problem of our research. See the literature for details: John G. Daugman, "High Confidence Visual Recognition of Persons by a Test of Statistical Independence," IEEE Transaction on Pattern Analysis and Machine Intelligence, volume15, no.11, pp.1148-1161, 1993. Iris image quality assessment is a very important link in the automatic iris recognition system, which ensures that the images that meet the quality standards are obtained during the acquisition process. In practice, due to the focal length of the acquisition device during shooting, the rotation of the eyeball at the moment of shooting, and the partial occlusion of the iris by the eyelids and eyelashes, it is often impossible to perform subsequent feature extraction on the collected iris image. An effective iris image quality evaluation model has not been proposed in the existing algorithms, so we aim to establish a general and feasible evaluation model, see the literature for details: Chen Ji, Hu Guangshu, "Iris Image Quality Evaluation based on Wavelet Packet Decomposition," Journal of Tsinghua University (Sci & Tech), volume 43, no.3, pp.377-380, 2003.

The commonly used iris positioning methods are:

(1) Two-step iris localization method based on gray gradient. It finds the approximate position of the inner and outer edges of the iris through rough positioning, and then uses a circular detector to perform fine positioning within a small range around this position, so as to find the precise position of the inner and outer edges of the iris. However, in practical applications, this method requires repeated iterative searches, which has a large amount of calculation and is not efficient. See literature for details: Li Qingrong, Ma Zheng, "A Iris Location Algorithm," Journal of UEST of China, volume 31, no.1, pp.7-9.

(2) Iris location method based on Hough transform. It uses a certain operator to extract the edge points in the iris image, so as to search for the position of the circular curve that passes the most edge points. Its disadvantage is that noise is often introduced in edge point extraction, which makes the result of iris location inaccurate. See the literature for details: Richard P. Wildes, "Iris Recognition: an Emerging Biometric Technology," Proceedings of the IEEE, volume85, pp.1348-1363, 1997.

The existing iris quality assessment methods are:

(1) Method based on fast Fourier transform. It performs two-dimensional fast Fourier transform on the pixels in two rectangular blocks on the iris area, and then analyzes whether the image is clear and has eyelash occlusion through the statistics of its high-frequency, intermediate-frequency and low-frequency energy. The generality of the model is not strong, and it is easy to misjudge a clear iris image with less texture as a low-quality iris image. See literature for details: Li Ma, Tieniu Tan, Yunhong Wang, Dexin Zhang, "Personal Identification based on Iris Texture Analysis," IEEE Transactions on Pattern Analysis and Machine Intelligence, volume.25, no.12, pp.1519-1533.

(2) The method based on wavelet packet decomposition. It selects the sub-band with the most concentrated distribution of texture high-frequency components as the characteristic sub-band, and uses its energy as the criterion for judging the image quality. The disadvantage of this method is that it cannot judge iris images which are problematic due to eyelash occlusion. See literature for details: Chen Ji, Hu Guangshu, "Iris Image Quality Evaluation based on Wavelet Packet Decomposition," Journal of Tsinghua University (Sci & Tech), volume 43, no.3, pp.377-380, 2003.

The above-mentioned iris image preprocessing algorithms all have problems to a certain extent. The positioning algorithm takes a lot of time, and is easily disturbed by the eyelash occlusion problem, and the stability is not high. The generality of iris image quality assessment method is not strong.

Contents of the invention

The task of the present invention is to provide an iris positioning method based on gray gradient and curve fitting, which has the characteristics of accurate positioning under the condition of eyelashes blocking. And on this basis, a set of iris image quality evaluation model with strong generality is established.

In order to describe the content of the present invention conveniently, some terms are defined first.

Definition 1: Iris. The center of the eyeball is the black pupil, and the ring-shaped tissue between the outer edges of the pupil is the iris. It presents interlaced texture features resembling spots, filaments, striations, and crypts. The iris of the same person hardly changes throughout a person's life, and the iris of different people is completely different.

Definition 2: Grayscale image. An image that contains only luminance information in the image without any other color information.

Definition 3: Binarization threshold. The grayscale threshold value used when binarizing the image.

Definition 4: Binarization. The process of converting all values of the entire image into only two values, generally these two values are 0 and 1 or 0 and 255. When the value on the image is greater than or equal to the threshold value of binarization, the value of the point is binarized to 1 (or 255); when the value on the image is smaller than the threshold value of binarization, the value of the point is binary into 0.

表式全包含，

为膨胀运算符，Θ为腐蚀运算符，o为开启运算符，g为闭合运算符。Definition 5: Mathematical Morphology. Use structural elements with a certain shape to measure and extract the corresponding shape in the image to achieve the purpose of image analysis and recognition. There are four basic operations in mathematical morphology: dilation (or expansion), corrosion (or erosion), opening and closing. The formulas for dilation and erosion are: $A\⊕B=\left\{x\right|{\left(\hat{B}\right)}_x\∩A\≠\mathrm{\φ}\}$ and $\mathrm{A\ΘB}=\left\{x\right|{\left(B\right)}_x\⊆A\};$ The calculation formula for the opening operation is: $\mathrm{AoB}=\left(\mathrm{A\ΘB}\right)\⊕B$ and $\mathrm{AgB}=(A\⊕B)\mathrm{\ΘB}.$ Among them, A is an image collection, B is a structural element, ^ means to map about the origin, () _x means translation x, ∩ means intersection, φ means empty set,

The form is all inclusive,

is the dilation operator, Θ is the corrosion operator, o is the opening operator, and g is the closing operator.

Definition 6: Grayscale projection. Project the grayscale in two-dimensional space to one-dimensional space, which is divided into horizontal grayscale projection and vertical grayscale projection. Horizontal grayscale projection refers to accumulating the grayscale in a two-dimensional image along the horizontal direction and converting it to a one-dimensional space. The conversion function is: $S_h\left(x\right)=\underset{\mathrm{the\; y}=1}{\overset{N}{\mathrm{\Σ}}}I(x,\mathrm{the\; y}).$ Vertical grayscale projection refers to accumulating the grayscale in the two-dimensional image along the vertical direction and converting it to a one-bit space. The conversion function is: $S_v\left(\mathrm{the\; y}\right)=\underset{x=1}{\overset{m}{\mathrm{\Σ}}}I(x,\mathrm{the\; y}).$ Among them, _Sh (x) represents the grayscale projection value whose abscissa is x, S _v (y) represents the grayscale projection value whose ordinate is y, M, N are the width and height of the image, and I(x, y) is The gray value of the pixel at position (x, y).

Definition 7: pupil boundary point. Refers to the point on the inner edge of the iris at the outer edge of the pupil.

Definition 8: Circle fitting. Knowing the coordinates of a series of points, establish a circular curve equation that best reflects the positions of these coordinate points. Specifically: the circle equation is x ² +y ² +cx+dy+e=0, c, d and e are parameters about the radius of the circular curve and the coordinate point of the center of the circle, and (x, y) is a point on the circular curve coordinate values, then the best circular curve relative to these coordinate points is to minimize the sum of error variances. The formula for the sum of error variances is: ${\mathrm{\&epsiv;}}^2=\underset{i}{\mathrm{\&Sigma;}}{(x_{\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="A20061012252800151.png"\; state="\{\{state\}\}">i</figure-callout>}}^2+{\mathrm{the\; y}}_i^2+{\mathrm{cx}}_i+{\mathrm{dy}}_i+e)}^2,$ Among them, ε ² refers to the sum of error variances, and ( _xi , y _i ) are the coordinates of known points.

Definition 9: Horizontal first-order difference. In the image, subtract the gray value of the front pixel from the gray value of the back pixel of a row, or subtract the gray value of the back pixel from the gray value of the front pixel to get the horizontal first-order difference value of the row . The horizontal first-order difference can highlight the vertical edge information of the image, which is convenient for edge extraction.

Definition 10: The boundary point of the outer edge of the iris. The iris is a ring-shaped area, and the points located on the outer edge of the iris are called iris outer border points.

Definition 11: Normalization. Stretch the circular iris area into a rectangular area of the same size to eliminate the influence of factors such as pupil constriction on the recognition effect due to different shooting distances. The specific calculation formula is: $\left\{\begin{array}{c}x(r,\mathrm{\&theta;})=(1-r)x_p\left(\mathrm{\&theta;}\right)+{\mathrm{r\; x}}_i\left(\mathrm{\&theta;}\right)\\ \mathrm{the\; y}(r,\mathrm{\&theta;})=(1-r){\mathrm{the\; y}}_p\left(\mathrm{\&theta;}\right)+{\mathrm{ry}}_i\left(\mathrm{\&theta;}\right)\end{array}\right.,$ Among them, r is distributed in the interval [0, 1], θ is distributed in the interval [0, 2π], and (x _p (θ), y _p (θ)) and ( _xi (θ), y _i (θ)) Denote the iris inner boundary point and outer boundary point in the θ direction, respectively.

Definition 12. Normalized iris image. The rectangular image obtained after normalizing the original iris image.

Definition 13.8 - Neighborhood. For a pixel p whose coordinate point is (x, y), it has 4 horizontal and vertical adjacent pixels, and their coordinates are (x+1, y), (x-1, y), (x , y+1), (x, y-1), and 4 diagonal adjacent pixels, their coordinate points are (x+1, y+1), (x+1, y-1), (x -1, y+1), (x-1, y-1), such 8 pixels are collectively called the 8-neighborhood of p.

Definition 14. Point sharpness. The operator used to evaluate the sharpness of digital images, its mathematical form is: $f=\underset{i=1}{\overset{m\&times;\mathrm{no}}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{8}{\mathrm{\&Sigma;}}}|\mathrm{iGO}/\mathrm{dx}|/m\&times;\mathrm{no},$ Among them, dI is the difference between the gray value of the pixel point in the eight neighborhoods of a certain point in the image and the gray value of the point, dx is the distance between adjacent points, m and n are the height and width of the image respectively.

Definition 15. Effective pixels. It refers to the pixels located in the iris area in the normalized iris image, mainly to distinguish invalid pixels such as eyelashes and eyelids. The conditions that valid pixels must meet are: V _lash ≤ I(x, y) ≤ V _eyelid , where V _lash and V _eyelid are thresholds for judging whether they are pixels located in the eyelash area and eyelid area, and I(x, y) is the grayscale of the image.

Definition 16. Visibility. Refers to the visibility of the iris texture in the original iris image, and the main factor affecting the visibility is the occlusion of the iris area by the eyelids and eyelashes.

According to the iris image preprocessing algorithm of the present invention, it comprises the following steps:

Step 1. The iris in the human eye is image-collected by a camera device to obtain an original grayscale image containing the iris image.

Step 2. Choose a fixed threshold value V _b , binarize the original iris image, assign the gray value of the pixel point whose gray value is greater than the threshold value V _b in the original gray image to 1, and assign the gray value of the pixel point smaller than the threshold value V _b The gray value of the pixel point is assigned as 0.

为膨胀运算符，Θ为腐蚀运算符。Step 3. For the binary image obtained in step 2, perform a closing operation in mathematical morphology to eliminate small holes in the binary image. Specifically, the closure operation is: $\mathrm{AgB}=(A\&CirclePlus;B)\mathrm{\&Theta;B},$ That is to say, the expansion operation is performed on the original image A with the structural element B, and then the erosion operation is performed. The structure element B is a 7×7 matrix, the values of the elements in the middle approximately circular area are 1, and the values of the other elements are 0. g is the closing operator,

is the dilation operator, and Θ is the erosion operator.

Step 4. Obtain the horizontal and vertical gray scale projection of image in the calculation step 3, the computing formula of horizontal projection is: $S_h\left(x\right)=\underset{\mathrm{the\; y}=1}{\overset{N}{\mathrm{\&Sigma;}}}I(x,\mathrm{the\; y}),$ The formula for calculating the vertical grayscale projection is: $S_v\left(\mathrm{the\; y}\right)=\underset{x=1}{\overset{m}{\mathrm{\&Sigma;}}}I(x,\mathrm{the\; y}).$ Among them, _Sh (x) represents the grayscale projection value whose abscissa is x, S _v (y) represents the grayscale projection value whose ordinate is y, M, N are the width and height of the image, and I(x, y) is The gray value of the pixel at position (x, y).

Step 5. Search for the abscissa _x o when the horizontal grayscale projection Sh (x) takes the minimum value in step 4 and the vertical coordinate y _o when the vertical grayscale projection _Sh (y ₎ takes the minimum value, set (x _o , y _o ) as the rough center of the pupil.

Step 6. On the line whose ordinate is y _o , take (x _o , y _o ) as the center, and search leftward along the horizontal direction for a point whose pixel gray value is greater than T, when the searched pixel gray value is greater than T Stop searching immediately when , record the coordinates (x _l , y _o ) at this time as the coordinates of the boundary point of the pupil, and then search to the right along the horizontal direction in the same way, and obtain the coordinates of another boundary point (x _r , y _o ).

Step 7. Take several lines near the coordinate point (x _o , y _o ), and search for the pupil boundary point on each line taken out. The method is the same as the search method in the line y _o in step 6, and finally you can get a Coordinates of series pupil boundary points.

Step 8. Since the inner edge of the pupil is very similar to a circle, perform circle fitting on a series of pupil boundary points obtained in step 7, specifically: the circle equation is x ² +y ² +cx+dy+e =0, c, d and e are parameters about the radius of the circular curve and the center coordinate point, (x, y) is the coordinate value of the point on the circular curve, so the best circular curve relative to these coordinate points is to make the error Variance and min. The formula for the sum of error variances is: ${\mathrm{\&epsiv;}}^2=\underset{i}{\mathrm{\&Sigma;}}{(x_{\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="A20061012252800151.png"\; state="\{\{state\}\}">i</figure-callout>}}^2+{\mathrm{the\; y}}_i^2+{\mathrm{cx}}_i+{\mathrm{dy}}_i+e)}^2,$ Among them, ε ² refers to the error variance sum, (x _i , y _i ) is the coordinates of known points, and finally the precise center (x _p , y _p ) and radius r _p of the pupil are obtained.

Step 9. Calculate the horizontal first-order difference of the row where the coordinate point (x _p , y _p ) is located. The specific calculation formula is: when x _p < x < N-5, D(x, y _p )=I(x+5, y _p )-I(x, y _p ); when 5<x≤x _p , D(x, y _p )=I(x-5, y _p )-I(x, y _p ). Among them, D(x, y _p ) represents the horizontal first-order difference value of the coordinate point (x, y _p ), I(x, y) represents the gray value of the coordinate point (x, y), and N is the image width.

Step 10. Calculate the sum of the horizontal first-order differences between each point and the next 20 points on the interval [x _p +r _p +20, x _p +r _p +100] on the line whose ordinate is y _p . The specific calculation formula is: when x _p +r _p +20<x<x _p +r _p +100, $S(x,{\mathrm{the\; y}}_p)=\underset{i=0}{\overset{20}{\mathrm{\&Sigma;}}}\mathrm{D.}(x+i,{\mathrm{the\; y}}_p),$ Where D(x+i, y _p ) is the horizontal first-order difference value of the coordinate point (x+i, y _p ) obtained in step 7. And in this interval, find the corresponding coordinate point ( _xi , y _p ) when S(x, y _p ) takes the maximum value as the boundary point of the outer edge of the iris.

Step 11. On the line whose ordinate is y _p , calculate the sum of the horizontal first-order differences between each point and the previous 20 points on the interval [x _p -r _p -100, x _p -r _p -20]. The specific calculation formula is: when x _p -r _p -100<x<x _p -r _p -20, $S(x,{\mathrm{the\; y}}_p)=\underset{i=0}{\overset{20}{\mathrm{\&Sigma;}}}\mathrm{D.}(x-i,{\mathrm{the\; y}}_p),$ Where D(x+i, y _p ) is the horizontal first-order difference value of the coordinate point (x+i, y _p ) obtained in step 9. And find the coordinate point (x, y _p ) corresponding to the maximum value of S(x, y _p ) in this interval as the boundary point of the outer edge of the iris.

Step 12. Get some lines near the coordinate point (x _p , y _p ), carry out the search of the iris outer edge boundary point on each line taken out, the method is carried out in the y _p line in step 9, step 10 and step 11 The search method is the same, and finally a series of coordinates of the iris outer edge boundary points can be obtained.

Step 13. Since the outer edge of the pupil is also very similar to a circle, a series of iris outer edge boundary points obtained in step 10 are subjected to a circle fitting similar to that in step 8 to obtain the precise center of the iris outer edge (x _i , y _i ) and radius r _i .

Step 14. Perform normalization processing on the located iris area, and the specific calculation formula is: $\left\{\begin{array}{c}x(r,\mathrm{\&theta;})=(1-r)x_p\left(\mathrm{\&theta;}\right)+{\mathrm{r\; x}}_i\left(\mathrm{\&theta;}\right)\\ \mathrm{the\; y}(r,\mathrm{\&theta;})=(1-r){\mathrm{the\; y}}_p\left(\mathrm{\&theta;}\right)+{\mathrm{ry}}_i\left(\mathrm{\&theta;}\right)\end{array}\right.$ Among them, r is distributed in the interval [0, 1], θ is distributed in the interval [0, 2π], and (x _p (θ), y _p (θ)) and ( _xi (θ), y _i (θ)) Denote the iris inner boundary point and outer boundary point in the θ direction, respectively.

Step 15. the point sharpness of the normalized iris image that arrives in the size that calculates step 14 is M * N, and concrete computing formula is: $f=\underset{i=1}{\overset{m\&times;\mathrm{no}}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{8}{\mathrm{\&Sigma;}}}|\mathrm{iGO}/\mathrm{dx}|/m\&times;\mathrm{no},$ Among them, dI is the difference between the gray value of the pixel point in the 8-neighborhood of a certain point in the image and the gray value of the point, dx is the distance between adjacent points, and m and n are the height and width of the image respectively.

Step 16: Compare the point sharpness value f of the normalized iris image obtained in step 15 with the preset threshold value V _f for judging whether the iris image is clear. If f≥V _f , the image is considered The clarity of the system meets the requirements of the system, otherwise, it is considered not to meet the requirements of the system.

Step 17. count the number of effective pixels in the normalized iris image, and the specific calculation formula is: $K=\underset{x}{\mathrm{\&Sigma;}}\underset{\mathrm{the\; y}}{\mathrm{\&Sigma;}}\mathrm{no}(x,\mathrm{the\; y}),$ in $\mathrm{no}(x,\mathrm{the\; y})=\left\{\begin{array}{cc}1,& V_{\mathrm{lash}}\&le;I(x,\mathrm{the\; y})\&le;V_{\mathrm{eyelid}}\\ 0,& I(x,\mathrm{the\; y})>V_{\mathrm{eyelid}}\mathrm{or\; I}(x,\mathrm{the\; y}){<V}_{\mathrm{lash}}\end{array}\right.,$ Among them, V _lash and V _eyelid are thresholds for judging whether they are pixels located in the eyelash area and the eyelid area, and I(x, y) is the grayscale of the image.

Step 18. Compare the number K of effective pixels obtained in step 17 with the preset threshold value V _k for judging whether there is eyelid and eyelash occlusion in the iris image. If K≥V _k , it is considered The visibility of the iris image meets the requirements of the system, otherwise, it is deemed not to meet the requirements of the system.

Through the above steps, we extract the normalized iris image from the original image containing iris, and judge whether the image meets the requirements of the system.

It should be noted:

1. Perform binarization in step 2 to select a fixed threshold V _b The selected fixed threshold V _b is obtained through a large number of experiments, and a fixed threshold is selected here because the gray value of the pupil area and the iris The gray value of the region is very different, even the iris image taken under different lighting conditions can guarantee the effect of binarization.

2. Locating the rough center of the pupil (x _o , y _o ) in step 5 is to determine the range of iris boundary point search.

3. In step 6, it is considered that the point whose gray value is greater than T is the boundary point of the pupil, because there is an obvious increase in the gray value at the edge of the pupil, and when it is greater than a certain value, it is the edge of the pupil.

4. In step 5, since the first-order difference horizontal projection curve obtained in step 4 has many burrs, it is not conducive to the precise positioning of the method, and the Gaussian function must be used to smooth the projection value.

5. It is mentioned in step 15 that the size of the normalized iris image is M×N, the value M is determined according to the interval value of θ taken during the normalization operation, and the value N is determined by the value of r taken during the normalization operation The interval value is determined.

6. The point sharpness value of the image in step 15 mainly represents the clarity of the image. The larger the point sharpness value f is, the clearer the image is, and the smaller f is, the blurrier the image is.

7. The threshold V _f in step 16 is obtained by testing a large number of iris images from the same acquisition device, and this value can accurately classify clear and fuzzy iris images.

8. V _lash and V _eyelid are mentioned in step 17. We believe that the pixel points with a gray value smaller than V _lash are pixels in the eyelash area, and those with a gray value greater than V _eyelid are pixels in the eyelid area. The gray value Distributed between V _lash and V _eyelid are pixels in the iris area.

9. The larger the mailbox pixel number K value obtained in step 17, the larger the unoccluded iris area, and the smaller the K value, the more serious the existing eyelid and eyelash occlusion.

The present invention adopts the combination of boundary point search and circle fitting, first realizes the rough positioning of the center of the pupil circle through binarization, corrosion, expansion and grayscale projection; then searches out the boundary through single grayscale value comparison and grayscale difference accumulation Points and curve fitting; Finally, according to the obtained normalized iris image, the image quality is evaluated from two aspects of clarity and visibility. Adopting the method based on the combination of boundary point search and curve fitting proposed by the present invention can effectively improve the iris positioning accuracy; adopting the quality evaluation method based on point sharpness and effective point number proposed by the present invention improves the traditional quality evaluation. Algorithm versatility.

The innovation of the present invention is:

Make full use of the grayscale information of the iris image and the method of curve fitting, and perform curve fitting by obtaining the coordinates of the boundary points of the inner and outer circles of the iris, so as to obtain the position information of the iris area and achieve the purpose of separating the iris; and through normalization The gray information of the iris image correctly evaluates the quality of the iris image. The present invention first adopts the projection method to carry out horizontal and vertical gray scale projection on the binary iris image filled with light spots to obtain the rough center of the pupil. By scanning the horizontal grayscale curve, a point whose grayscale value is greater than a certain threshold is found as the inner boundary point of the iris. Circle fitting is performed on the above series of inner boundary points, so as to obtain the position information of the inner edge of the iris. Afterwards, by integrating the horizontal first-order difference of 20 adjacent points, the position coordinates when the integral value takes the maximum value are obtained as the boundary points of the iris, and then these boundary points are used for circle fitting to obtain the outer edge of the iris location information. It is a characteristic of the present invention to use the method of combining grayscale information and circle fitting to locate the iris area. Compared with the general two-step iris positioning method, the positioning accuracy of the present invention is 5 percentage points higher, and the speed is improved. 60%. Then when performing quality assessment, the present invention statistically normalizes the point sharpness and the number of effective pixel points of the iris image, and by comparing with the preset threshold value, the iris image quality has been correctly evaluated, and Very versatile.

Description of drawings

Figure 1 is the original image containing the iris;

Among them, 1 represents the pupil; 2 represents the iris; 3 represents the spot in the pupil; 4 represents the inner edge of the iris; 5 represents the outer edge of the iris.

Fig. 2 is a schematic diagram of the 8-neighborhood of the pixel p;

where r is the horizontal and vertical adjacent pixels, and s is the diagonal adjacent pixels.

Fig. 3 is a diagram of the positioning result of the method of the present invention.

Detailed ways

Adopt method of the present invention, at first use C language and assembly language to write iris preprocessing program; Then adopt CMOS or CCD camera device to take the original image of iris automatically; Then the iris original image that is taken is input to DSP embedded system as source data It is processed in the iris preprocessing program; after iris positioning and image quality evaluation, the qualified iris image is positioned and output iris normalized image containing rich texture information. Using 2,400 gray-scale iris images of different people with different lighting conditions and different shooting postures as source data, the positioning accuracy rate is 97.5%, and it only takes 100ms to locate an image.

In summary, the method of the present invention makes full use of the grayscale information of the iris image, combined with the circle fitting method, so as to quickly and accurately locate the iris region from the provided iris original image and make an accurate quality assessment.

Claims (2)

1. the present invention relates to a kind of iris recognition preprocessing method based on gray scale information, it is characterized in that comprising the steps: Step 1. Through the camera device, the iris in the human eye is image-acquired to obtain the original grayscale image containing the iris image; Step 2. Select a fixed threshold V _b to binarize the original iris image; Step 3. to the binary image obtained in step 2, carry out the closing operation in the mathematical morphology to eliminate the small hole in the binary image; Step 4. Calculate the horizontal pair projection _Sh (x) and the vertical grayscale projection _Sv (y) of the image obtained in step 3; and search for the abscissa x _o when the horizontal grayscale projection _Sh (x) takes the minimum value And the vertical grayscale projection _Sh (y) takes the minimum value as the vertical coordinate y _o , and (x _o , y _o ) is regarded as the rough center of the pupil; Step 5. On the line where the point (x _o , y _o ) is located and several lines near the point, search for a series of coordinates of pupil boundary points by comparing the gray values; Step 6. Since the inner edge of the pupil is very similar to a circle, a series of pupil boundary points obtained in step 5 are circle-fitted, and finally the precise center of circle (x _p , y _p ) and radius r _p of the pupil are obtained; Step 7. Calculate the horizontal first-order difference of the row where the coordinate point (x _p , y _p ) is located; and on the row whose ordinate is y _p and its adjacent rows, search for the horizontal first-order difference value in the interval that may exist on the outer edge of the iris The corresponding coordinate points when the sum takes the maximum value can get a series of iris outer edge boundary points; Step 8. Since the outer edge of the pupil is also very similar to a circle, therefore, a series of iris outer edge boundary points obtained in step 7 are fitted with a circle similar to that in step 6 to obtain the precise center of the iris outer edge (x _i , y _i ) and radius r _i ; Step 9. Normalize the located iris region; Step 10. Calculate the point sharpness of the normalized iris image obtained in step 11, and compare it with the preset threshold value V _f for judging whether the iris image is clear. If f≥V _f , the image is considered The clarity of the system meets the requirements of the system, otherwise, it is deemed not to meet the requirements of the system; Step 11. Statistically normalize the number of effective pixels in the iris image, and compare the number K of effective pixels with the preset threshold value V _k for judging whether the iris image has the problem of eyelid and eyelash occlusion In comparison, if K≥V _k , it is considered that the visibility of the iris image meets the requirements of the system; otherwise, it is considered not to meet the requirements of the system; Through the above steps, we can extract the normalized iris image from the original image containing iris, and judge whether the image meets the requirements of the system. 2. as claimed in claim 1, a kind of iris recognition preprocessing method based on grayscale information, is characterized in that: Step 5. On the line where the ordinate is y _o , with (x _o , y _o ) as the center, search for a point with a pixel gray value greater than T along the horizontal direction to the left, when the searched pixel gray value is greater than T Stop searching immediately when , record the coordinates (x _l , y _o ) at this time as the coordinates of the boundary point of the pupil, and then search to the right along the horizontal direction in the same way, and obtain the coordinates of another boundary point (x _r , y _o ); take several lines near the coordinate point (x _o , y _o ), and search for pupil boundary points on each line taken out, the method is the same as the search method in y _o line, and finally a series of pupil boundaries can be obtained the coordinates of the point; Step 7. (1) Calculate the horizontal first-order difference of the row where the coordinate point (x _p , y _p ) is located. The specific calculation formula is: when x _p <x<N-5, D(x, y _p )=I( x+5, y _p )-I(x, y _p ); when 5<x≤x _p , D(x, y _p )=I(x-5, y _p )-I(x, y _p ); Wherein D(x, y _p ) represents the horizontal first-order difference value of the coordinate point (x, y _p ), I(x, y) represents the gray value of the coordinate point (x, y), and N is the image width; (2) Calculate the sum of the horizontal first-order differences between each point and the next 20 points on the interval [x _p +r _p +20, x _p +r _p +100] on the line whose ordinate is y _p ; The specific calculation formula is: when x _p +r _p +20<x<x _p +r _p +100, $S(x,{\mathrm{the\; y}}_p)=\underset{i=0}{\overset{20}{\mathrm{\&Sigma;}}}\mathrm{D.}(x+i,{\mathrm{the\; y}}_p),$ Among them, D(x+i, y _p ) is the horizontal first-order difference value of the coordinate point (x+i, y _p ) obtained in step (1); and find out S(x, y _p ) in this interval to take The coordinate point corresponding to the maximum value is used as the boundary point of the outer edge of the iris; then calculate the horizontal first-order difference value between each point and the previous 20 points on the interval [x _p -r _p -100, x _p -r _p -20] The sum; the specific calculation formula is: when x _p -r _p -100<x<x _p -r _p -20, $S(x,{\mathrm{the\; y}}_p)=\underset{i=0}{\overset{20}{\mathrm{\&Sigma;}}}\mathrm{D.}(x-i,{\mathrm{the\; y}}_p),$ Among them, D(xi, y _p ) is _the horizontal first-order difference value of the coordinate point (xi, y _p ) obtained in step (1); The coordinate point of is used as the boundary point of the outer edge of the iris; (3) Take several lines near the center of the pupil (x _p , y _p ), and search for the boundary points of the outer edge of the iris on each line taken out. The method is the same as the search method for the line y _p in step 7. Finally, it can be Obtain the coordinates of a series of iris outer edge boundary points; Step 10. the point sharpness of the normalized iris image that arrives in the size that calculates in step 9 is M * N, and concrete computing formula is: $f=\underset{i=1}{\overset{m\&times;\mathrm{no}}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{8}{\mathrm{\&Sigma;}}}|\mathrm{iGO}/\mathrm{dx}|/m\&times;\mathrm{no},$ Wherein dI is the difference between the gray value of the pixel point in the 8-neighborhood of a certain point in the image and the gray value of the point, dx is the distance between adjacent points, and m and n are respectively the height and width of the image; Step 11: Compare the point sharpness value f of the normalized iris image obtained in step 9 with the preset threshold value V _f for judging whether the iris image is clear. If f≥V _f , the image is considered The clarity of the system meets the requirements of the system, otherwise, it is considered not to meet the requirements of the system.

Images