CN103020612A - Device and method for acquiring iris images - Google Patents

Abstract

The invention discloses a device and a method for acquiring iris images. The device comprises an imaging module, an illuminating module, a three-dimensional movement platform, a control module, an optical platform, fixing rods and a jaw support, wherein the jaw support is fixed onto the optical platform via the fixing rods, and the imaging module is fixed onto the three-dimensional movement platform; the illuminating module comprises light source fixing plates, a light intensity regulator and near-infrared LEDs, the near-infrared LEDs are arranged on the light source fixing plates annularly, and infrared rays emitted by the near-infrared LEDs are parallel to a primary optical axis of a CCD (charge coupled device) camera; and the three-dimensional movement platform comprises a lead screw and slider mechanism, a stepper motor and a driver. The device and the method have the advantages that a portion corresponding to an eye of a person can be photographed in the middle of an image by adjusting the device according to the location of the eye of the person, the left eye and the right eye of the person can be photographed by controlling the three-dimensional movement platform as needed when the three-dimensional movement platform moves, quick centering and focusing can be realized by an automatic centering and focusing algorithm, and the clear iris images can be photographed.

Description

A kind of iris image acquiring device and method

Technical field

The invention discloses a kind of non-reaction type, contactless automatic iris image acquiring equipment, more specifically say a kind of iris image acquiring device and method.

Background technology

Iris is as important identity diagnostic characteristics, has the advantages such as untouchable, uniqueness, stability, non-infringement.Identity identification system based on iris mainly is comprised of four parts: iris image acquisition, image pre-service, iris feature extract, matching and recognition.Obtaining of iris image is the first step of iris recognition, is a very important link.Yet because the diameter of iris is very little, and the iris image that gathers requires that abundant pixel must be arranged, so obtaining of iris image also is a very difficult link.Because that existing iris collection device still exists is expensive, the person of being verified is had requirement, needs the problems such as manually-operated, therefore await further development research.Along with the progress of technology, cheap, fast and easy, contactless automatic iris collection device become main direction of studying.

Utilize the Iris Location method to determine the iris inner boundary, in more existing documents, provided detailed description, such as Ye Yongqiang, Shen Jianxin, week makes a whistling sound. based on the quick Iris Location [J] of pupil gray feature. photoelectric project, 2010,37 (3): 127 ~ 132; Ye Yongqiang. the location of iris image and feature extraction, [master thesis]. Nanjing: Nanjing Aero-Space University, 2010.

Summary of the invention

The invention provides and a kind ofly can realize clear iris image acquiring device and the method for obtaining iris image.

For solving the problems of the technologies described above, a kind of iris image acquiring device of the present invention, comprise image-forming module, lighting module, three-dimensional motion platform, control module, optical table, fixed bar and jaw holder, described jaw holder is fixed on the optical table by fixed bar, described image-forming module is fixed on the three-dimensional motion platform, it comprises tight shot, CCD camera and semi-transparent semi-reflecting lens, and wherein, semi-transparent semi-reflecting lens is between CCD camera and tight shot; Described lighting module comprises light source fixed head, dimmer and is arranged on the near-infrared LED luminotron of circular array on the light source fixed head, and the infrared beam of described near-infrared LED luminotron is parallel with the primary optical axis of CCD camera; Described three-dimensional motion platform comprises screw mandrel slide block mechanism, stepper motor and driver.

Further, in the iris image acquiring device of the present invention, the focal length of described tight shot is 35mm.

Further, in the iris image acquiring device of the present invention, described screw mandrel slide block mechanism comprises grinding screw rod, guide rail and mobile platform.

Further, in the iris image acquiring device of the present invention, described CCD camera is near infrared black-white CCD camera.

Further, in the iris image acquiring device of the present invention, the quantity of described near-infrared LED luminotron be 4-6 and equal angles be arranged on the light source fixed head.

Simultaneously, in order to realize the automatic centering focusing, the present invention also proposes a kind of iris image acquiring method, may further comprise the steps:

Step 1, at first CCD camera position initialization, drive motor is gone to the zero point of setting automatically by the CCD camera, and this zero point, per 4000 pulses of motor are set a pause, so that CCD collected by camera image away from focusing;

Step 2, utilize Iris-orientation Algorithm to determine the iris inner boundary;

Step 3, according to the iris inner boundary of determining in the step 2, utilize focus algorithm to obtain the evaluation of estimate of iris image;

Step 4, judge the iris image acquiring number of times, when times of collection greater than 1 the time, then the evaluation of estimate with upper width of cloth image compares, and judges whether evaluation of estimate increases, if step 1 is then returned in the evaluation of estimate increase, otherwise execution in step 5; When times of collection less than or equal to 1 the time, return execution in step 1;

Step 5, judge the iris image acquiring number of times, when times of collection greater than 2 the time, execution in step 6; Otherwise return step 1;

Step 6, judge whether evaluation of estimate reduces continuously, if reduce continuously then execution in step 7, otherwise return step 1;

Step 7, stepper motor is segmented control, whether judge the segmentation number of times less than 2, if less than 2 step-length is reduced 1/2, will segment afterwards number of times and add 1 and order about the CCD camera and move round about, and return step 1, otherwise execution in step 8;

Step 8, the corresponding point of evaluation of estimate of getting last collection image and adjacent 2 points thereof utilize these three points of conic fitting and calculate intermediate point, and mobile CCD camera to intermediate point is finished focusing and execution in step 9 afterwards;

The clear iris image of focusing is finished in step 9, collection;

Step 10, iris inner circle center, location are determined iris inner circle center and picture centre difference DELTA X and the Δ Y on x axle and y axle, and are judged Δ X ²+ Δ Y ²Whether less than 100, if less than 100 end operations, move Δ X otherwise drive the three-dimensional motion platform along directions X, move Δ Y along Y-direction, and return step 9.

Further preferred version in a kind of iris image acquiring focusing method of the present invention, utilizes focus algorithm to determine the evaluation of estimate of iris image in the step 3, be specially,

The shade of gray value of (3-1) choosing any five points of calculating iris inner boundary is as evaluation criterion, and the coordinate of each point is:

$\left(\begin{array}{c}x\\ y\end{array}\right)=r\left(\begin{array}{c}\cos \left(\mathrm{\θ}\right)\\ \sin \left(\mathrm{\θ}\right)\end{array}\right)+\left(\begin{array}{c}{x}_0\\ y_0\end{array}\right)$

In the formula, x ₀, y ₀The coordinate in the presentation video pupil center of circle; θ represents the point of lower boundary and the angle of interpupillary line and zero degree radius; R represents the size of pupil radius;

(3-2) on iris inner boundary both sides on the radial direction every N(N=2,3) individual pixel gets a gray-scale value compute gradient, gets altogether 25 points, obtain respectively the M value of those 25 feature pixels and get wherein maximum value:

G _n=max(M ₁,M ₂,M ₃,....M ₂₅)

(3-3) determine the evaluation of estimate of iris image, P is as follows for its evaluation of estimate:

$p=\underset{n=1}{\overset{T}{\mathrm{\Σ}}}\left(G_n\right|G_n>C)/T$

In the formula, C is the threshold value of a shade of gray of setting, and T meets the requirements of number.

The present invention compared with prior art has significant advantage:

1, designed a three-dimensional motion platform, can be according to the position adjustment at eyes place so that eyes be full of the image middle, and can be as required, the movement of control three-dimensional motion platform realizes the shooting of right and left eyes.Automatic centering focus algorithm in the application controls software can realize fast centering focusing, the iris image of shooting clear.

2, adopt near-infrared light source, can effectively avoid the unintelligible phenomenon of iris in normal natural light or daylight lamp situation, extracted, also can prevent owing to adopting the excessive eyes sense of discomfort that causes of visible light light intensity.

3, use semi-transparent semi-reflecting lens, can effectively eliminate the reflection spot of visible light on the iris image, reduce picture noise.

4, adopt the black-white CCD camera, effectively solved color camera to the unconspicuous characteristics of near infrared light sensitization, make the image texture of collection clear, contrast is strong, reduces the identification error rate.

5, device cheap, need not human assistance, noncontact.

6, adopt the shade of gray evaluation map image sharpness of iris inner boundary to finish the iris focusing, compare calculated amount and greatly reduce with Measurement for Digital Image Definition commonly used such as energy gradient function method, frequency spectrum function method, Entropy Function Method, arithmetic speed is obviously accelerated, and is conducive to accelerate the picking rate of iris image.

The present invention is described in further detail below in conjunction with the drawings and specific embodiments;

Description of drawings

Fig. 1 is iris image acquiring device one-piece construction figure;

Fig. 2 is the lighting device floor map;

Fig. 3 is three-dimensional motion platform schematic diagram;

Fig. 4 (a) is subjective formula lighting circuit schematic diagram; Fig. 4 (b) objective type lighting circuit schematic diagram;

Fig. 5 is iris image acquiring focusing process flow diagram;

Fig. 6 is system's focusing running block scheme;

Embodiment

Such as Fig. 1, Fig. 2, shown in Figure 3, the iris image acquiring device of a kind of automatic centering focusing of the present invention, comprise image-forming module, lighting module, three-dimensional motion platform, control module 9, optical table 7, fixed bar and jaw holder 5, described jaw holder is fixed on the optical table by fixed bar, described image-forming module is fixed on the three-dimensional motion platform 6, it comprises tight shot 3, CCD camera 1 and semi-transparent semi-reflecting lens 2, and wherein, semi-transparent semi-reflecting lens is between CCD camera and tight shot; Described lighting module comprises light source fixed head 8, dimmer and is arranged on the near-infrared LED luminotron 4 of circular array on the light source fixed head, and the infrared beam of described near-infrared LED luminotron is parallel with the primary optical axis of CCD camera; Described three-dimensional motion platform comprises screw mandrel slide block mechanism, stepper motor 61 and driver; Described screw mandrel slide block mechanism comprises grinding screw rod 62, guide rail 63 and mobile platform 64.

(1) image-forming module

Image-forming module is comprised of three parts such as tight shot of the near infrared black-white CCD camera of 945nm sensitivity, semi-transparent semi-reflecting lens, focal distance f=35mm.According to the size of human eye, focusing is from greatly about about 200mm; Semi-transparent semi-reflecting lens is between CCD camera and tight shot; Wherein the CCD camera has the asynchronous collecting function, this function can capture the high quality graphic that does not have litura under very high shutter speed, can eliminate in shooting process owing to rocking of photographer caused impact for picture quality, the valid pixel scope of camera is 752 * 582, satisfies pixel request.

(2) lighting module

Lighting module be by 4 to 6 near-infrared LED light source equal angles be installed on the light source fixed head, its direction and camera lens, CCD main shaft are parallel; The wavelength of light source is 940nm, and brightness can be regulated.The mode of regulating is divided into two kinds: subjective formula and objective type.Subjective formula refers to that the staff relies on the bright-dark degree of eye-observation image to regulate the light-source brightness size, such as Fig. 4 (a); Objective type refers to by Computer nomogram image brightness and draws analyze conclusion then according to conclusion brightness adjusting size, such as Fig. 4 (b).

(3) three-dimensional motion module

The three-dimensional motion module mainly is comprised of screw mandrel slide block mechanism, precision stepper motor, the driver that three direction ranges are respectively 30mm, 30mm, 150mm.Adopt grinding screw rod, helical pitch is 1mm; The accurate spherical guide that intersects, material of main part is aluminium alloy; The resolution of precision stepper motor is 0.25 μ (20 segmentation), loses step≤2 μ.The maximum translational speed of module can reach 10mm/sec, repetitive positioning accuracy 5 μ, absolute fix precision 10 μ.

(4) control module

Control module mainly is comprised of control computing machine, software algorithm and relevant interface.

Such as Fig. 5, shown in Figure 6, a kind of iris image acquiring focusing method specifically may further comprise the steps:

Step 1, CCD position initialization at first, drive motor is gone to the zero point of setting automatically by CCD, and this zero point, per 4000 pulses of motor are set a pause away from focusing, so that CCD gathers image;

Step 2, utilize Iris-orientation Algorithm to determine the iris inner boundary, be specially:

Setting threshold is with the iris image binaryzation, pixel less than threshold value is 1, pixel greater than threshold value is 0, obtain bianry image, it is white only having the color of pupil and eyelashes part this moment among the figure, utilize the centroid detection method to be partitioned into pupil region for the point of quality is arranged with white portion among the figure is abstract, centroid detection method concrete steps are as follows:

(1) supposes that at first the binary picture geometric center is barycenter (x ₀, y ₀);

(2) centered by the barycenter of hypothesis, take out larger rectangle subgraph A ₁' (i, j) calculates the barycenter of subgraph

(3) with new barycenter Centered by, take out less square subgraph A ₂' (i, j) calculates the barycenter of subgraph

(4) calculate the side-play amount (Δ x, Δ y) of barycenter in former figure,

(5) whether judge Δ x, Δ y all less than the threshold value of setting, if just finish, if otherwise jump to step (3);

Just can get rid of the interference of eyelashes through iteration, search out pupil center, comprise the iris inner boundary in the last square subgraph that takes out, search iris internal boundary points in this square, inner boundary shape approximation circle is so utilize at last circular function to adopt least square method to come match iris inner boundary.

Step 3, according to the iris inner boundary of determining in the step 2, utilize focus algorithm to obtain the evaluation of estimate of iris image, be specially:

The shade of gray value of (3-1) choosing any five points of calculating iris inner boundary is as evaluation criterion, and the coordinate of each point is:

$\left(\begin{array}{c}x\\ y\end{array}\right)=r\left(\begin{array}{c}\cos \left(\mathrm{\θ}\right)\\ \sin \left(\mathrm{\θ}\right)\end{array}\right)+\left(\begin{array}{c}{x}_0\\ y_0\end{array}\right)$

In the formula, x ₀, y ₀The coordinate in the presentation video pupil center of circle; θ represents the point of lower boundary and the angle of interpupillary line and zero degree radius; R represents the size of pupil radius;

(3-2) on iris inner boundary both sides on the radial direction every N(N=2,3) individual pixel gets a gray-scale value compute gradient, gets altogether 25 points, obtain respectively the M value of those 25 feature pixels and get wherein maximum value:

G _n=max(M ₁,M ₂,M ₃，....M ₂₅)

(3-3) determine the evaluation of estimate of iris image, its evaluation of estimate is as follows:

$p=\underset{n=1}{\overset{T}{\mathrm{\Σ}}}\left(G_n\right|G_n>C)/T$

In the formula, C is the threshold value of a shade of gray of setting, and T meets the requirements of number;

Step 4, judge the iris image acquiring number of times, when times of collection greater than 1 the time, then the evaluation of estimate with upper width of cloth image compares, and judges whether evaluation of estimate increases, if step 1 is then returned in the evaluation of estimate increase, otherwise execution in step 5; When times of collection less than or equal to 1 the time, return execution in step 1;

Step 5, judge the iris image acquiring number of times, when times of collection greater than 2 the time, execution in step 6; Otherwise return step 1;

Step 6, judge whether evaluation of estimate reduces continuously, if reduce continuously then execution in step 7, otherwise return step 1;

Step 7, stepper motor is segmented control, whether judge the segmentation number of times less than 2, if less than 2 step-length is reduced 1/2, will segment afterwards number of times and add 1 and order about CCD and move round about, and return step 1, otherwise execution in step 8;

Step 8, the corresponding point of evaluation of estimate of getting last collection image and adjacent 2 points thereof utilize these three points of conic fitting and calculate intermediate point, and mobile CCD finishes to intermediate point and focuses and execution in step 9 afterwards;

The clear iris image of focusing is finished in step 9, collection;

Step 10, iris inner circle center, location are determined difference DELTA X and the Δ Y of iris inner circle center and picture centre, and are judged Δ X ²+ Δ Y ²Whether less than 100, if less than 100 end operations, move Δ X otherwise drive the three-dimensional motion platform along directions X, move Δ Y along Y-direction, and return step 9.

Claims (7)

1. iris image acquiring device, it is characterized in that: comprise image-forming module, lighting module, three-dimensional motion platform, control module, optical table, fixed bar and jaw holder, described jaw holder is fixed on the optical table by fixed bar, described image-forming module is fixed on the three-dimensional motion platform, this image-forming module comprises tight shot, CCD camera and semi-transparent semi-reflecting lens, wherein, semi-transparent semi-reflecting lens is between CCD camera and tight shot; Described lighting module comprises light source fixed head, dimmer and is arranged on the near-infrared LED luminotron of circular array on the light source fixed head, and the infrared beam of described near-infrared LED luminotron is parallel with the primary optical axis of CCD camera; Described three-dimensional motion platform comprises screw mandrel slide block mechanism, stepper motor and driver.

2. iris image acquiring device according to claim 1, it is characterized in that: the focal length of described tight shot is 35mm.

3. iris image acquiring device according to claim 1, it is characterized in that: described screw mandrel slide block mechanism comprises grinding screw rod, guide rail and mobile platform.

4. iris image acquiring device according to claim 1, it is characterized in that: described CCD camera is near infrared black-white CCD camera.

5. iris image acquiring device according to claim 1 is characterized in that: the quantity of described near-infrared LED luminotron be 4-6 and equal angles be arranged on the light source fixed head.

6. one kind is utilized that the described iris image acquiring device of any one claim carries out iris image acquiring method in the claim 1 to 5, it is characterized in that, specifically may further comprise the steps:

Step 1, at first CCD camera position initialization, drive motor is gone to the zero point of setting automatically by the CCD camera, and this zero point, per 4000 pulses of motor are set a pause, so that CCD collected by camera image away from focusing;

Step 2, utilize Iris-orientation Algorithm to determine the iris inner boundary;

Step 3, according to the iris inner boundary of determining in the step 2, utilize focus algorithm to obtain the evaluation of estimate of iris image;

Step 4, judge the iris image acquiring number of times, when times of collection greater than 1 the time, then the evaluation of estimate with upper width of cloth image compares, and judges whether evaluation of estimate increases, if step 1 is then returned in the evaluation of estimate increase, otherwise execution in step 5; When times of collection less than or equal to 1 the time, return execution in step 1;

Step 5, judge the iris image acquiring number of times, when times of collection greater than 2 the time, execution in step 6; Otherwise return step 1;

Step 6, judge whether evaluation of estimate reduces continuously, if reduce continuously then execution in step 7, otherwise return step 1;

Step 7, stepper motor is segmented control, whether judge the segmentation number of times less than 2, if less than 2 step-length is reduced 1/2, will segment afterwards number of times and add 1 and order about the CCD camera and move round about, and return step 1, otherwise execution in step 8;

Step 8, the corresponding point of evaluation of estimate of getting last collection image and adjacent 2 points thereof utilize these three points of conic fitting and calculate intermediate point, and mobile CCD camera to intermediate point is finished focusing and execution in step 9 afterwards;

The clear iris image of focusing is finished in step 9, collection;

Step 10, iris inner circle center, location are determined iris inner circle center and picture centre difference DELTA X and the Δ Y on x axle and y axle, and are judged Δ X ²+ Δ Y ²Whether less than 100, if less than 100 end operations, move Δ X otherwise drive the three-dimensional motion platform along directions X, move Δ Y along Y-direction, and return step 9.

7. iris image acquiring method according to claim 6 is characterized in that: utilizes focus algorithm to determine the evaluation of estimate of iris image in the step 3, is specially,

The shade of gray value of (3-1) choosing any five points of calculating iris inner boundary is as evaluation criterion, and the coordinate of each point is:

$\left(\begin{array}{c}x\\ y\end{array}\right)=r\left(\begin{array}{c}\cos \left(\mathrm{\θ}\right)\\ \sin \left(\mathrm{\θ}\right)\end{array}\right)+\left(\begin{array}{c}{x}_0\\ y_0\end{array}\right)$

In the formula, x ₀, y ₀The coordinate in the presentation video pupil center of circle; θ represents the point of lower boundary and the angle of interpupillary line and zero degree radius; R represents the size of pupil radius;

(3-2) on radial direction, get a gray-scale value compute gradient every N pixel on iris inner boundary both sides, wherein, N=2,3, get altogether 25 points, obtain respectively the shade of gray value M value of those 25 feature pixels and get wherein maximum value:

G _n=max(M ₁,M ₂,M ₃,....M ₂₅)

(3-3) determine the evaluation of estimate of iris image, p is as follows for its evaluation of estimate:

$p=\underset{n=1}{\overset{T}{\mathrm{\Σ}}}\left(G_n\right|G_n>C)/T$

In the formula, C is the threshold value of a shade of gray of setting, and T meets the requirements of number.

Images