WO2013087026A1

WO2013087026A1 - Locating method and locating device for iris

Info

Publication number: WO2013087026A1
Application number: PCT/CN2012/086665
Authority: WO
Inventors: 张祥德; 王琪; 单成坤; 周军
Original assignee: 北京天诚盛业科技有限公司
Priority date: 2011-12-16
Filing date: 2012-12-14
Publication date: 2013-06-20
Also published as: CN102902970A; CN102521576A

Abstract

The present invention relates to a locating method and locating device for the iris. The method includes the following steps: step 1, performing facula detection on an iris image; step 2, judging whether the iris image swings; step 3, initially locating the pupils; and step 4, locating the iris. The beneficial effects of the present invention are: locating the edges inside the iris using a 2D circular Gabor filter and a circular weight calculus detection operator, deleting the noise and exception points on the external boundary using a combined method of the circular weight calculus detection operator and clustering to obtain correct external boundary locating, thus reducing the interference to the iris identification system by the noise and exception condition, and improving the stability and accuracy of the iris locating algorithm.

Description

Iris positioning method and positioning device

TECHNICAL FIELD The present invention relates to the field of iris localization, and in particular to an iris positioning method and a positioning device based on a two-dimensional circular Gabor filter and a circular weighted calculus detection operator. BACKGROUND OF THE INVENTION With the development of information technology, human beings have entered an unprecedented information age. Information security has also become a hot issue that people are paying more and more attention to. In daily life, we often need to verify the identity of ourselves or others. Reliable identification can make our lives avoid troubles and protect the security of personal information. Reliable identification technology also makes financial and business transactions safer and more effective. . Traditional identity authentication is more and more difficult to meet the needs of society because it is extremely easy to forge and lose. With the development of biological science, biometrics has attracted more and more people's attention because of its high security and reliability. , has become a cross-cutting topic in the study of applied mathematics, image processing and pattern recognition. Biometrics are combined by high-tech means such as computers, biosensors, and biostatistics to identify and certify a person's identity using the inherent physiological and behavioral characteristics of the person. At present, common biometric recognition technologies mainly include fingerprint recognition, iris recognition, face recognition and speech recognition. The international iris recognition market has maintained rapid growth and is expected to reach $1,481 million in revenue by 2015. Iris recognition has good prospects and has been widely used in customs, airports, banking, access control and information security. As iris recognition technology matures, it will play an increasingly important role in protecting people's information security. The iris is a circular pigmented membrane around the pupil of the eye. The iris-rich texture pattern is the basis of iris recognition. Iris recognition technology extracts, analyzes and matches these textures. The uniqueness and stability of the iris texture make the iris an excellent biometric for human identification. The iris recognition system first collects the iris image, then divides the iris region in the acquired iris image, and finally extracts features on the normalized iris image for matching, and compares the similarities of the two iris images to determine whether they are from the same personal. The iris is a ring-shaped tissue that is outside the pupil (black part of the eye) and inside the sclera (white part of the eye). Compared with other biometrics, the iris has many obvious advantages as an identification feature, including:

1. High uniqueness: The iris has highly unique and complex texture features, the iris texture has more than one hundred degrees of freedom, and there are almost no irises with the same texture features; 2. Good stability: The iris is formed in the embryonic stage. Due to the protection of the cornea, the iris is less affected by the external environment and the texture features are stable.

3. Good anti-counterfeiting: The zooming characteristics of the iris with the change of light intensity and its regular vibration, provide a natural test method for forged iris, it is almost impossible to change the texture of the iris; 4. Non-invasive: iris The image is acquired without physical contact and can be obtained within a certain distance without causing discomfort to the human body. Iris positioning is a key part of the iris recognition system. The accuracy of positioning is related to the feature extraction and matching of the iris. Therefore, the accuracy of the iris positioning results directly affects the accuracy and efficiency of the iris recognition system. The human eye image obtained by the iris image acquisition system includes the eyelids, eyelashes, sclera, pupils, etc. of the human eye. The iris positioning is a process of detecting the inner and outer boundaries of the iris and separating the iris from other parts of the human eye. The iris has a good ring structure, and from the pupil to the iris to the sclera, the gray scale of the image rises stepwise. This good edge structure facilitates the positioning and segmentation of the iris. However, the eyelids and eyelashes may block the iris area. During the specific iris positioning process, proper pre-processing can eliminate interference and improve the stability and accuracy of the overall positioning algorithm. In addition, during image acquisition, the size of the pupil changes with the illumination environment. The eyelids and eyelashes block the iris area. The distance between the human eye and the acquisition device is also uncertain, and there is still a certain angle between them. Relationship, so even for different images of the same person's eyes, the positioning results are different. In some classic iris recognition systems, the iris boundary is approximated by a circular shape, and the inner and outer boundaries of the iris are located using a non-concentric double circle model. Since the shape is known, the corresponding circular parameters can be obtained according to the boundary characteristics of the iris. In view of the problem of poor accuracy of low-quality iris positioning in the prior art, an effective solution has not yet been proposed. SUMMARY OF THE INVENTION The present invention provides a new iris positioning method and positioning device. The positioning method and the positioning device are based on iris positioning of a two-dimensional circular Gabor filter and a circular weighted calculus detection operator, and the iris image is determined. The interference of the noise information on the iris recognition system is reduced, and the stability and accuracy of the iris localization algorithm are improved. The object of the present invention is achieved by the following technical solutions: An iris positioning method includes the following steps: Step 1. The spot detection of the iris image specifically includes the following sub-steps: Step 1.1: Perform image acquisition on the iris in the human eye through the camera device, and smooth the obtained iris image by using a median filtering method; Step 1.2: Filter the smoothed iris image by using a two-dimensional circular Gabor filter Action, then select the binarization threshold, use the binarization method to determine the spot area; Step 2, determine whether the iris image is shaken: For the spot area detected in step 1.2, use the least squares ellipse fitting to calculate the spot area boundary The determined ellipse, if the ratio of the ellipse length to the half axis is greater than 1.55, the iris image is sloshing, otherwise it is a non-sloshing iris image; Step 3, initial positioning of the pupil, specifically including the following substeps: Step 3.1, step 2 The obtained clear iris image is reduced to 0.2 times; Step 3.2: The reduced iris image in step 3.1 is smoothed by the median filtering method; Step 3.3, 3.2 results, using a two-dimensional circular Gabor filter Filtering, the coordinate of the maximum value of the filtering result is used as the estimated center of the pupil; Step 4 Positioning the iris specifically includes the following sub-steps: Step 4.1: Using the parabolic weighted calculus detection operator to position the upper and lower eyelids of the iris image obtained in step 3.2; Step 4.2, returning the iris image obtained in step 3.2 The gradient value normalized by the iris image is subtracted from the normalized gray value, and the difference is greater than 0.1 for the eyelash region, and less than 0.1 for the non-eyelash region; Step 4.3, removing the iris image obtained in step 3.2 Eyelashes and eyelids in the middle; Step 4.4. Using the circular weighted calculus detection operator to locate the inner boundary of the iris for the image obtained in step 4.3; Step 4.5: Positioning the outer boundary of the iris: Calculating the circular weighted calculus of the outer boundary Sub-value, taking the largest 10 of all values as possible candidate boundaries. Then, using the clustering method, iteratively excludes the noise points in the boundary points, and calculates the circular weighted calculus detection operator value after eliminating the noise point, and takes the boundary parameter corresponding to the maximum gradient after eliminating the noise point as the outer boundary of the iris. Position the results. In the step 1.2, the smoothed iris image is convoluted by a two-dimensional circular Gabor filter, and the maximum value is the position of the spot, and the maximum value of the convolution result M_max is greater than 0.6*M_max. For the spot area. The median filtering method is a 5*5 median filtering method. According to another aspect of the present invention, there is provided an iris positioning apparatus for performing any of the iris positioning methods provided by the present invention described above. According to another aspect of the present invention, an iris positioning device is provided, the device comprising: a detecting module for detecting a spot of an iris image; a determining module for determining whether an iris image is shaken; and a first positioning module Initial positioning of the pupil; and a second positioning module for positioning the iris. Further, the detecting module includes: a collecting sub-module, configured to perform image capturing on the iris in the human eye by the imaging device; and a first smoothing processing sub-module, configured to perform smoothing processing on the collected iris image by using a median filtering method; The filtering sub-module is configured to filter the smoothed iris image by using a two-dimensional circular Gabor filter; and a binarization sub-module for selecting a binarization threshold, and determining a spot area by using a binarization method. Further, the determining module includes: a calculating sub-module, configured to calculate, by using a least square method, a ellipsoid determined by a boundary of the spot area, and a determining sub-module for determining a ratio of the length of the ellipse to the half-axis of the ellipse; Among them, if the ratio of the ellipse length to the half axis is greater than 1.55, the iris image is swaying, otherwise it is a clearer iris image. Further, the first positioning module includes: a scaling sub-module for reducing a clear iris image to an original 0.2 times; a second smoothing processing sub-module for using the median filtering method to reduce the reduced iris image Performing a smoothing process; and determining a sub-module for filtering the smoothed image by using a two-dimensional circular Gabor filter, and taking the coordinate of the maximum value of the filtered result as the estimated center of the pupil. Further, the second positioning module comprises: an eyelid positioning sub-module, configured to position the upper and lower eyelids by using the parabola to obtain the iris image obtained by the second smoothing processing sub-module; and the eyelash positioning sub-module is configured to obtain the second smoothing processing sub-module The iris image is normalized, and the normalized gradient value of the iris image is subtracted from the normalized gray value, and the difference is greater than 0.1 for the eyelash region, less than 0.1 for the non-eyelash region; a module for removing eyelashes and eyelids in an iris image obtained by the second smoothing processing sub-module; and a boundary positioning sub-module for positioning the inner and outer boundaries of the iris image obtained by deleting the sub-module by using a weighted calculus operator . By the present invention, an iris positioning method comprising the steps of: detecting a spot of an iris image; determining whether the iris image is shaken; initial positioning of the pupil; and positioning the iris, since determining the spot area before performing iris positioning And determine whether the iris image is shaking. When performing iris positioning, only The non-sloshing iris image is used for iris positioning, so that the iris positioning accuracy caused by the positioning of the shaking iris image can be avoided, and the method of iris positioning only for the non-sloshing iris image improves the accuracy of iris positioning. The invention has the following beneficial effects: the two-dimensional circular Gabor filter and the circular weighted calculus detection operator are used to locate the inner boundary of the iris, thereby reducing the interference of the noise information on the iris recognition system, and the iris can be quickly and accurately located. The inner boundary of the image, in the outer boundary of the iris, uses the clustering method to eliminate noise and abnormal pixels, which improves the stability and accuracy of the iris localization algorithm. DRAWINGS

The accompanying drawings, which are incorporated in the claims of the claims 1 is a flow chart of an iris positioning method according to an embodiment of the present invention; FIG. 2 is a schematic diagram of an image containing an iris; FIG. 3 is a schematic diagram of a two-dimensional circular Gabor filter; 4 is a schematic diagram showing the result of filtering the smoothed iris image by a two-dimensional circular Gabor filter; FIG. 5 is a schematic diagram showing the result of the spot region detection; FIG. 6 is a schematic diagram showing the result of reducing and smoothing the iris image; Figure 8 is a schematic diagram showing the results of the positioning of the eyelashes; Figure 9 is a schematic diagram of the results of the iris positioning; Figure 10 is a schematic block diagram of an iris positioning device according to an embodiment of the present invention; A schematic block diagram of an iris positioning device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS It should be noted that the embodiments in the present application and the features in the embodiments may be combined with each other without conflict. The invention will be described in detail below with reference to the drawings in conjunction with the embodiments. In order to conveniently describe the present invention, some terms are first defined. Definition 1: Iris. The center of the eye is a black pupil, and the annular tissue between the outer edges of the pupil is the iris, which presents interlaced texture features similar to spots, filaments, stripes, and crypts. The iris of the same person will hardly change in a person's life. The iris of different people is completely different. Definition 2: Binarization threshold. The grayscale threshold used when binarizing the image. Definition 3: Binarization. Converting all values of the entire image into a process with only two values, typically two values of 0 and 1 or 0 and 255. When the value on the image is greater than or equal to the binarization threshold, the value of the point is binarized to 1 (or 255); when the value on the image is less than the binarization threshold, the value of the point is binarized to 0. . Definition 4: Least squares ellipse fit. Firstly, the ellipse-based algebraic distance-based least squares method is used to obtain the preliminary estimate of the ellipse parameter. Then, the geometric distance-based least squares method is used to iterate to give the optimal estimate of the ellipse parameter, and combined with the detection of the sphere, Constraints on multiple elliptical parameters achieve high detection accuracy. The ellipse fitting method based on least squares is applicable to various complex object models, and can intuitively give a measure of a certain fitting error, achieving a high fitting accuracy. Definition 5: Median filtering. The median filtering method is a nonlinear processing method that suppresses noise. For a given n values { al, a2, ..., an}, they are arranged in order of size. When n is an odd number, the value at the middle position is called the median of the n values. When n is an even number, the average of the two values at the intermediate position is referred to as the median of the n values. The output of a pixel after image median filtering is equal to the median value of the gray level of each pixel in the pixel. Definition 6: Parabolic weighted calculus detection operator. Let the parabolic equation be y = a(x - bf + c, find the parameter b, c) that makes the value of the following formula the largest _:

That is, the parabola obtained by arg max v ia, b, c) is taken as the final detection result < where, the gray value of the pixel point ( , x) is represented. c{ ^ ^i^fe indicates parabola

J

J = S(X - b) ² + C, the weighted curve integral of the gray value / y, x), N is the effective parabolic dispersion The number of pixels. In the discrete state, it is the weighted summation of the gray values of the pixel points on the circumference. G c) is a smoothing function, using a normal distribution function, which is a weight, which varies with s. Definition 7: Circular weighted calculus detection operator. Let the equation of the circle be ϊ - ; ^2: + ^: - X _s := , find the parameter that gives the largest value of the following formula (y ₀ , x ₀ , r ₀ ) t

Request arg max v _c (y _l

I(y, x)

Where /(_y,x) represents the gray value of the pixel point (_y, x). W _r . indicates that (j., x.) is the center, r. On the circumference of the radius, the weighted curve of the gray value / (_y, x) is integrated, and in the discrete state, the weighted sum of the gray values of the pixel points on the circumference. G r) is a smoothing function, using a normal distribution function, ^ is a weight, varies with s. Definition 8: Two-dimensional circular Gabor filter filtering. The two-dimensional circular Gabor filter function used in this embodiment is as follows:

G (x, y) = cos (ω (χ - χ ₀ ) ) β Let the image be / C3⁄4j), then the above filter function convolves the image, namely:

The above process is to perform two-dimensional circular Gabor filtering on the image f( '), which is the obtained two-dimensional Gabor filtering result. As shown in FIG. 1 , an iris positioning method according to an embodiment of the present invention includes the following steps: after collecting an iris in a human eye, performing spot detection on the collected iris image, determining a spot area, and then determining according to the determined The spot area determines whether the iris image is shaken. When the iris image is shaken, no processing is performed. The step ends. When the iris image is a non-sloshing iris image, the iris is initially positioned, that is, the position of the pupil is determined, and finally, according to the determined The pupil position locates the iris and, when performing iris positioning, includes detection of the eyelashes and eyelids and positioning of the iris boundary. According to the iris positioning method provided by the embodiment, before the iris positioning, the spot area is first determined, and the iris image is determined to be shaken. When the iris is positioned, only the non-sloshing iris image is iris-positioned, thereby avoiding shaking the iris. The accuracy of iris positioning caused by image positioning is poor, and the method of iris positioning only for non-sloshing iris images improves the accuracy of iris positioning. Preferably, the iris positioning method provided by this embodiment can be implemented by the following preferred steps. Step 1. Detecting the spot of the iris image, specifically including the following sub-steps: Step 1.1: Perform image acquisition on the iris in the human eye through the camera device, and the acquired iris image is as shown in FIG. 2, using a median filtering method. The obtained iris image is smoothed; Step 1.2: The smoothed iris image is filtered by a two-dimensional circular Gabor filter (as shown in FIG. 3), and the filtering result is shown in FIG. 4, and then the binarization threshold is selected. The binarization method is used to determine the spot area, and the image of the obtained spot area is as shown in FIG. 5; Step 2. Determine whether the iris image is shaken: Calculate the spot area detected in step 1.2 by least squares ellipse fitting calculation The ellipse determined by the boundary of the spot area, if the ratio of the length of the ellipse to the long axis is large, the iris image is swaying, otherwise it is a non-sloshing iris image; it should be noted that, after counting, 3000 images of the iris image obtained by normal shooting are The ratio of the length of the ellipse to the edge of the spot is about 1 to 1.30; the edge of the spot in the iris image is swayed. The ratio of the length of the ellipse to the length of the ellipse is greater than 1.80. Therefore, when judging whether the iris image is shaken, the threshold value can be determined by setting a threshold value, and the threshold value can be any value in the interval [1.30, 1.80], preferably, When the threshold is selected to be 1.55, that is, when the ratio of the ellipse length to the half axis is greater than 1.55, it is determined that the iris image is swaying. This method is also effective for iris images taken by other devices, and the threshold can be set by the same experiment. Step 3: Initially positioning the pupil, specifically including the following sub-steps: Step 3.1, reducing the clear iris image obtained in step 2 to 0.2 times, wherein when step 2 determines that the iris image is a shaking iris image The iris image is blurred due to shaking. When the iris image is a non-sloshing iris image, the iris image is relatively clear, and thus the clear iris image in step 3.1 is a non-sloshing iris image; Step 3.2: Smoothing the reduced iris image in step 3.1 by using the median filtering method, and the obtained iris image is as shown in FIG. 6; Step 3.3, the result obtained in step 3.2, that is, the smoothed iris image is used. The two-dimensional circular Gabor filter is filtered again, and the maximum value of the filtering result is taken to perform initial positioning of the pupil, and the position of the pupil is determined. Preferably, the coordinate of the maximum value of the filtering result is used as the estimation center of the pupil; Step 4 Positioning the iris specifically includes the following sub-steps: Step 4.1: Positioning the iris image obtained in step 3.2, using the parabolic weighted calculus detection operator to position the upper and lower eyelids, and the eyelid positioning result is as shown in FIG. 7; Determine the parabola parameter according to the estimated center of the pupil determined in step 3.3, and set the estimated pupil center to ( _XQ , y _Q ). To get the position of the parabola quickly, set the function of the parabola to y _Q = a(x- Xo) ² +b, which reduces the search range of the parameters. Step 4.2: Normalize the iris image obtained in step 3.2, and subtract the normalized gradient value of the iris image from the normalized gray value. The difference is greater than 0.1 for the eyelash region, and less than 0.1 for the non-eyelash. The area, the lash area positioning result is shown in Figure 8; Step 4.3, removing the eyelashes and eyelids in the iris image obtained in step 3.2; Step 4.4, using the weighted calculus operator to the iris image obtained in step 4. 3) The inner and outer boundaries are positioned. Step 4.4 specifically includes: using the circular weighted calculus detection operator to locate the inner boundary of the iris for the image obtained in step 4.3, and then using the circular weighted calculus detection operator to calculate the outer boundary of the iris image obtained in step 4.3) The weighted calculus detection operator value is taken as the boundary corresponding to the N largest circular weighted calculus detection operator values, and the noise point and other abnormal points are respectively deleted by clustering, and then the circular weighted micro after the exclusion point is calculated. The integral detection operator value, and finally takes the boundary corresponding to the maximum value of the N values as the final result of the outer boundary position of the iris, wherein the abnormal points include: a spot, an eyelash, an eye boundary, an upper and lower eyelid, and the like. Preferably, in the outer boundary positioning, among the total circular weighted calculus detection operator values, the largest 10 {G, I. = 1 _; 2 i0. For each one, proceed as follows: Take the non-noise partial pixel point set ί ^·, ^·: on the corresponding iris boundary. make

Calculate the grayscale mean _3⁄4 and the grayscale variance σ _ί of the pixels in the F\ collection _:

Here, i i refers to the number of elements in the set.

(3) Let ^; = {( ^yy): fix^y^e O -\.5δ, +1.5S]}.

(4) Calculate the distance between the gray value of these points and the standard gray value for all elements in the set e:

Dl / (w' - 3^,. When d > 1, it means / (χ _φ 3^: not the same class as the sample set, 0^, 3) is removed from the 1^ set; when 4≤1, The representation /(x _i; , y _ijf ;) is the same class as the sample set and does not change anything.

(5) Repeat steps (2)-(4) until the change is no longer performed, and the screening process of the boundary points ends.

(6) Mark all the boundary points deleted in the above process as noise points. Then, in 10 cases, the weighted calculus operator values of all non-noise points are calculated as j = 1, 2...10.

10

Replacement page (Article 26) Finally, the iris outer boundary parameter d^^i^i;) corresponding to the optimal value W = wi x 3⁄4} is taken as the localization result of the outer boundary of the iris. The final positioning result of the iris is shown in Fig. 9. The step 1.2 is specifically: convolving the smoothed iris image by using a two-dimensional circular Gabor filter, where the maximum value is the position of the spot, and the maximum value of the convolution result M_max is greater than k*M_max. Part of it is the spot area. The median filtering method is a 5*5 median filtering method, where k is a dimensionless coefficient, and k is preferably in the range of 0.1 to 1, more preferably k = 0.6. The embodiment of the present invention further provides an iris positioning device. The iris positioning device provided by the embodiment of the present invention is introduced below. It should be noted that the iris positioning method in the embodiment of the present invention can be performed by the iris positioning device provided by the embodiment of the present invention, and the iris positioning device in the embodiment of the present invention can also be used to execute the iris provided by the embodiment of the present invention. Positioning method.

FIG. 10 is a schematic block diagram of an iris positioning device according to an embodiment of the present invention. As shown in FIG. 10, the positioning device includes a detecting module, a determining module, a first positioning module, and a second positioning module. The detecting module is configured to detect a spot of the iris image to obtain a spot area; the determining module is configured to determine whether the iris image is shaken according to the determined spot area, and when the iris image is a non-sloshing iris image, the first positioning module is used to The pupil is initially positioned to determine the position of the pupil, and the second positioning module is used to position the iris according to the position of the pupil to determine the boundary of the iris. With the iris positioning device provided by the embodiment, the determination module determines whether the iris image is shaken, and when the first positioning module and the second positioning module perform iris positioning, only the non-sloating iris image is iris-positioned, thereby avoiding shaking the iris The accuracy of iris positioning caused by image positioning is poor, and the method of iris positioning only for non-sloshing iris images improves the accuracy of iris positioning. 11 is a schematic block diagram of another iris positioning device according to an embodiment of the present invention. As shown in FIG. 11, the positioning device includes a detecting module, a determining module, a first positioning module, and a second positioning module, where the detecting module includes collecting a submodule, a first smoothing submodule, a filtering submodule, and a binarization submodule, the judging module includes a computing submodule and a judging submodule, and the first positioning module includes a scaling submodule, a second smoothing processing submodule, and a determining submodule The second positioning module includes an eyelid positioning sub-module, an eyelash positioning sub-module, a deletion sub-module, and a boundary positioning sub-module. The collecting sub-module is used for image capturing of the iris in the human eye by the camera device to obtain an iris image; the first smoothing processing sub-module is used for smoothing the collected iris image by using a median filtering method; the filtering sub-module is used for The smoothed iris image is filtered by a two-dimensional circular Gabor filter; the binarization sub-module is used to select the binarization threshold, and the binarization method is used to determine the spot area. The calculation sub-module is used for the spot area detected by the detection module, and the least-squares method is used to fit and calculate the ellipse determined by the boundary of the spot area; the judgment sub-module is used to determine the iris by determining whether the ratio of the ellipse length and the half-axis is greater than a preset value. Whether the image is a shaking iris image, wherein if the ratio of the length of the ellipse to the long axis is larger than the preset value, the iris image is swaying, otherwise it is a clearer, ie, non-sloshing iris image, where The preset value is preferably 1.55. The scaling sub-module is used to reduce the clearer iris image to the original 0.2 times; the second smoothing processing sub-module is used to smooth the reduced iris image by the median filtering method; The dimension circular Gabor filter filters the image processed by the second smoothing processing sub-module, and performs initial positioning of the pupil according to the filtering result, and determines the position of the pupil. Specifically, the coordinate of the maximum value of the filtered result is taken as the pupil. Estimate the center. The eyelid positioning sub-module is configured to locate the upper and lower eyelids by using the parabola to obtain the iris image obtained by the second smoothing processing sub-module; the eyelash positioning sub-module is used for normalizing the iris image obtained by the second smoothing processing sub-module, and the iris image is The normalized gradient value is subtracted from the normalized gradation value, and the result of the subtraction is compared with 0.1, the lash area is greater than 0.1, and the non-lash area is less than 0.1. The sub-module is configured to remove the eyelashes and the eyelids in the iris image obtained by the second smoothing processing sub-module; the boundary positioning sub-module is configured to locate the inner and outer boundaries of the iris image obtained by the deleting sub-module by using the weighted calculus operator. Preferably, the boundary positioning sub-module comprises an inner boundary positioning unit and an outer boundary positioning unit. The inner boundary locating unit locates the inner boundary of the iris by using the circular weighted calculus detection operator to delete the image obtained by the submodule. The outer boundary locating unit uses the circular weighted calculus detection operator to calculate the circular weighted calculus detection operator value for the outer boundary of the iris image obtained by the deletion submodule, and takes the N largest circular weighted calculus detection operator values. Corresponding boundaries are used to delete the noise points and other abnormal points respectively, and then calculate the circular weighted calculus detection operator value after the exclusion point, and finally take the boundary corresponding to the maximum value of the N values as the outer boundary of the iris. The final result. The value of the N value can be set according to the calculation capability of the positioning device and the calculation accuracy. Preferably, N is 10, and the outer boundary positioning method described in the method embodiment is used, and details are not described herein again. It should be noted that the steps shown in the flowchart of the accompanying drawings may be performed in a computer system such as a set of computer executable instructions, and, although the logical order is shown in the flowchart, in some cases, The steps shown or described may be performed in an order different than that herein. Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in the storage device by the computing device, or they may be separately fabricated into respective integrated circuit modules. Blocks, or a plurality of modules or steps in them, are implemented as a single integrated circuit module. Thus, the invention is not limited to any specific combination of hardware and software. The above is only the preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Claims

The claim request method and an iris positioning method are characterized in that the method comprises the following steps:

Step 1: detecting the spot of the iris image;

Step 2: Determine whether the iris image is shaken;

Step 3: Initial positioning of the pupil;

Step 4: Position the iris. The iris positioning method according to claim 1, wherein the step 1 further comprises: Step 1. 1): performing image acquisition on the iris in the human eye through the imaging device, and using the median filtering method to obtain the image The iris image is smoothed;

Steps 1. 2): The smoothed iris image is filtered by a two-dimensional circular Gabor filter, and then the binarization threshold is selected, and the spot area is determined by binarization. The iris positioning method according to claim 2, wherein in the step 2, the ellipse determined by the boundary of the spot area is calculated by a least square method for the spot area detected in the step 1.2), If the ratio of the length of the ellipse to the long axis is large, the iris image is swaying, otherwise it is a clearer iris image. 2倍; Steps: Step 2: The step of the clear iris image obtained in step 2 is reduced to the original 0.2 times; 3. 2): Use the median filtering method to smooth the reduced iris image in step 3. 1); Step 3. 3): Use the two-dimensional circular Gabor filter to take the step 3. 2. The maximum filtering result The value is the initial position of the pupil to determine the position of the pupil. The iris positioning method according to claim 4, wherein the step 4 further comprises: step 4. 1): positioning the upper and lower eyelids by using the parabola to obtain the iris image obtained in the step 3.2. 2): Normalize the iris image obtained in step 3. 2), subtract the normalized gradient value of the iris image from the normalized gray value, and subtract the result with 0.1约为非睫毛的范围内。 Compared with the eyelash area, less than 0.1 is a non-lash area;

Step 4. 3): removing the eyelashes and eyelids in the iris image obtained in step 3. 2); Step 4. 4): Use the weighted calculus operator to locate the inner and outer boundaries of the iris image obtained in step 4. 3), and accurately identify the iris image whose inner and outer boundaries are not circular. The iris positioning method according to claim 3, wherein if the ratio of the elliptical length to the minor axis is greater than 1.55, the iris image is shaken. The iris positioning method according to claim 6, wherein the step 3 further comprises: step 3.1): reducing the clear iris image obtained in step 2 to 0.2 times; step 3.2): utilizing The median filtering method smoothes the reduced iris image in step 3.1); Step 3.3): The result obtained in step 3.2 is filtered by a two-dimensional circular Gabor filter, and the coordinate of the maximum value of the filtering result is taken as the estimated center of the pupil. The iris positioning method according to claim 7, wherein the step 4 further comprises: step 4.1): locating the upper and lower eyelids of the iris image obtained in step 3.2) by using a parabolic weighted calculus detection operator. ;

Step 4.2): Normalize the iris image obtained in step 3.2), and subtract the normalized gradient value of the iris image from the normalized gray value. The difference is greater than 0.1 for the eyelash region, and less than 0.1 is Non-lash area;

Step 4.3): Remove the eyelashes and eyelids in the iris image obtained in step 3.2); Step 4.4): For the image obtained in step 4.3), use the circular weighted calculus detection operator to locate the inner boundary of the iris;

Step 4.5): Calculate the circular weighted calculus detection operator value by using the circular weighted calculus detection operator on the outer boundary of the iris image obtained in step 4.3), and take the 10 largest circular weighted calculus detection operator values. Corresponding boundaries are used to delete the noise points and other abnormal points respectively, and then calculate the circular weighted calculus detection operator value after the exclusion point, and finally take the boundary corresponding to the maximum value of the 10 values as the outer boundary of the iris. The final result. The iris positioning method according to any one of claims 2-8, wherein: in the step 1. 2), the smoothed iris image is convoluted by a two-dimensional circular Gabor filter, The large value point is the position of the spot, and the maximum value of the convolution result is M_max, and the part larger than k*M_max is considered to be the spot area. The iris positioning method according to claim 9, wherein k is 0.6. An iris positioning device, comprising:

a detecting module, configured to detect a spot of the iris image;

a judging module, configured to determine whether the iris image is shaken;

a first positioning module for initial positioning of the pupil;

The second positioning module is configured to position the iris. The iris positioning device according to claim 11, wherein the detecting module comprises: a collecting sub-module, configured to perform image capturing on an iris in a human eye by an imaging device; and a first smoothing processing sub-module, The collected iris image is smoothed by a median filtering method;

a filtering submodule for filtering the smoothed iris image using a two-dimensional circular Gabor filter;

The binarization sub-module is used to select the binarization threshold, and the binarization method is used to determine the spot area. The iris positioning device according to claim 12, wherein the determining module comprises: a calculating sub-module configured to calculate a spot area detected by the detecting module by using a least squares method to calculate a spot area boundary Ellipse;

The judging sub-module is configured to determine a ratio of the ellipse length and the minor axis, wherein if the ratio of the ellipse length to the semi-axis is greater than 1.55, the iris image is swaying, otherwise it is a clear iris image. The second positioning module is configured to reduce the sharper iris image to the original 0.2 times; the second smoothing a processing sub-module for smoothing the reduced iris image by a median filtering method;

The determining sub-module is configured to filter the smoothed image by using a two-dimensional circular Gabor filter, and take the coordinate of the maximum value of the filtered result as the estimated center of the pupil. The iris positioning device according to claim 14, wherein the second positioning module comprises: an eyelid positioning sub-module for positioning the iris image obtained by the second smoothing processing sub-module with a parabola Eyelid The eyelash locating sub-module, wherein the iris image obtained by the second smoothing sub-module is normalized, and the normalized gradient value of the iris image is subtracted from the normalized gray value, and the difference is greater than 0.1. The area of the eyelashes, less than 0.1 is a non-lash area;

Deleting a sub-module for removing eyelashes and eyelids in the iris image obtained by the second smoothing sub-module;

The boundary locating sub-module is configured to locate the inner and outer boundaries of the iris image obtained by the deleting sub-module by using a weighted calculus operator.