JP4345741B2 - Personal identification device - Google Patents

Description

The present invention, human ecological features, those about the personal identification equipment for identifying the individuals in particular using the characteristics of the face and eye of the iris (iris).

  The iris in the human eyeball is composed of muscles that adjust the degree of opening of the pupil around the pupil of the black eye, and wrinkles or the like that are generated radially outward from the pupil form an iris pattern. Iris is used as an ecological feature for personal identification because it is almost unchanged throughout its life after being completed in early childhood, and its iris pattern varies from person to person.

  2 is an external view showing an example of a conventional personal identification device, and FIG. 3 is a block diagram of FIG.

  As shown in FIG. 2, in the conventional personal identification device, one eye image camera 1 is installed in order to photograph the identified person's iris, and a host computer 10 is connected to the eye image camera 1. The host computer 10 includes an image capture unit 11 that converts an analog video signal output from the eye image camera 1 into a digital video signal by analog / digital (hereinafter referred to as “A / D”) conversion. 11 is used to identify an individual based on the digital video signal converted at step S11.

  As shown in FIG. 3, an iris processing unit 12 that processes a digital video signal output from the image capture unit 11 is provided in the host computer 10. The iris processing unit 12 includes an iris coding processing unit 12a that extracts an iris region from the transmitted digital video signal and creates an iris code S12a. The iris registration processing unit 12b, the iris dictionary 12c, And the iris collation process part 12d is connected. The iris registration processing unit 12b registers the iris code S12a in the iris dictionary 12c. The iris verification processing unit 12d performs personal identification by comparing the iris code S12a output from the iris coding processing unit 12a with the iris code registered in the iris dictionary 12c, and outputs the identification result. .

  Next, a personal identification method using the conventional personal identification device shown in FIGS. 2 and 3 will be described.

  In the conventional personal identification device, the eye image camera 1 photographs the iris of the person to be identified (user) and sends this analog video signal to the image capture unit 11 in the host computer 10. The image capture unit 11 captures an analog video signal sent from the eye image camera 1, converts the analog video signal into a digital video signal by A / D conversion, and converts the eye image data converted into the digital video signal to the iris processing unit 12. Send to.

  The iris processing unit 12 performs registration processing or collation processing of the iris code S12a using the transmitted eye image data. That is, in the iris processing unit 12, first, the iris coding processing unit 12a processes the eye image from the transmitted eye image to create an iris code S12a. At the time of registration, the created iris code S12a is registered in the iris dictionary 12c together with the personal information of the identified person in the iris registration processing unit 12b. On the other hand, at the time of collation, the iris collation processing unit 12d compares the generated iris code S12a with the iris dictionary 12c to perform collation processing for personal identification, and outputs the processing result.

JP-T 8-504979

However, the conventional personal identification equipment, in order to stably shoot the iris region, it is necessary to photograph the iris in front of the eye image camera 1. For this reason, in order to correctly photograph the iris with one eye image camera 1, any height identification person (that is, an identification person who is too high or too low) is placed in front of the eye imaging camera 1. In order to turn the iris, there is a case where it is necessary to take an unreasonable posture, which causes a burden of forcing the identified person to take a special posture.

  The present invention solves the problems of the prior art and provides a personal identification device that can correctly perform iris photography for any height person in a natural posture without forcing the person to be identified to be unreasonable. The purpose is to do.

In order to solve the above-mentioned problem, in the invention of claim 1 of the present invention, in the personal identification device, one or more rows are arranged in the vertical direction which is the height direction of the identified person, and each identified person is A plurality of eye image cameras that shoot and output a plurality of identified person images; a plurality of sensors that are installed beside each of the plurality of eye image cameras and that detect distances to the identified persons ; and Eye image determination processing means for determining and selecting an eye image from which the iris of the eye is most likely to be captured from the image of the identified person, estimating the height of the identified person from the detection result of the sensor, from each captured the plurality of the identified person image in the plurality of eye images camera, from the eye to the eye image roughly selected that appears to be reflected, the selection result by the eye image determining process unit, The eye is copied Select the eye image that appears to be determined to have been configured to perform personal identification using the iris pattern in the eye image.

  By adopting such a configuration, the person to be identified is photographed by a plurality of eye image cameras, the distance to the person to be identified is detected by a plurality of sensors, and the identification result of the person to be identified is detected from the detection results of these sensors. The height of the eye is estimated, and an eye image that is considered to have an eye is selected from a plurality of photographed images of the person to be identified, and personal identification is performed using an iris pattern in the eye image.

  According to a second aspect of the present invention, in the first aspect of the invention, a one-way transmissive cold mirror is disposed between the identified person and the plurality of eye image cameras in the longitudinal direction which is the height direction of the identified person. ing.

  According to the first aspect of the present invention, since the person to be identified is photographed by the plurality of eye image cameras, it is possible to take a natural posture without forcing an unreasonable posture to the person to be identified having different heights. Even tall people can do iris photography correctly. In addition, a plurality of identified person images taken by a plurality of eye image cameras are installed next to the plurality of eye image cameras, the distance to the identified person is detected, the height of the identified person is estimated. Therefore, since the eye image that is considered to have an eye is selected, the target eye image can be determined at high speed and with high accuracy, and the processing can be performed at high speed and with high accuracy.

  According to the second aspect of the present invention, since the cold mirror that is unidirectionally transparent is disposed between the person to be identified and the plurality of eye image cameras, the person to be identified is not conscious of the eye image camera, and the cold mirror The eye iris can be photographed by any one of the eye image cameras simply by photographing his face straight.

  The personal identification device is arranged in one or a plurality of rows in the vertical direction, which is the height direction of the identified person, a plurality of eye image cameras that respectively photograph the identified person and output a plurality of identified person images, and these And a plurality of sensors that are installed next to the plurality of eye image cameras and detect the distance to the person to be identified. Then, the height of the identified person is estimated from the detection result of the sensor, and an eye image that is considered to be in the eye is selected from a plurality of identified person images respectively captured by a plurality of eye image cameras. Individual identification is performed using an iris pattern in the image.

(Reference Example 1)
(Constitution)
4 is an external view of a personal identification device showing Reference Example 1 of the present invention, FIG. 5 is a configuration diagram of FIG. 4, and FIGS. 6A and 6B are configurations of the frame switcher 20 and the image capture unit 31 of FIG. 7 and 7 are configuration diagrams of the eye image determination processing unit 32 in FIG.

  As shown in FIG. 4, in the personal identification device of Reference Example 1, for example, a unidirectionally transmissive cold mirror 13 is installed, and an identified person 14 stands on the right side of the cold mirror 13. . On the left side of the cold mirror 13, a plurality of photographing means (for example, eye image cameras) 16-1 to 16-n are arranged in, for example, one row along the vertical direction that is the height direction of the person to be identified 14. Yes. Each of the eye image cameras 16-1 to 16-n captures the identified person 14 and outputs each identified person image S16-1 to S16-n composed of an analog video signal. The person to be identified 14 is not conscious of the eye image cameras 16-1 to 16-n, and simply displays his / her face straight on the cold mirror 13, and any one of the eye image cameras 16-1 to 16-n is used. Eye iris is being shot.

  A divided image generating means (for example, a frame switcher) 20 is connected to the plurality of eye image cameras 16-1 to 16-n, and a host computer 30 is connected to the frame switcher 20. The frame switcher 20 collectively divides a plurality of identified person images S16-1 to S16-n sent from a plurality of eye image cameras 16-1 to 16-n into one frame (one frame). An image S23 is generated, and the divided image S23 is sent to the host computer 30.

  A capture unit (for example, an image capture unit) 31 is provided in the host computer 30. The image capture unit 31 captures a single-frame multiple-divided image S23 that is sent, and outputs a single-frame multiple-divided image S31b composed of a digital video signal by A / D conversion. Then, the host computer 30 selects a desired eye image that appears to have an eye from the multiple divided images S31b output from the image capture unit 31, and registers an iris pattern in the eye image, or It has a function of performing personal identification by comparing a registered iris pattern with a newly input iris pattern.

As shown in FIG. 5, in the host computer 30, in addition to the image capture unit 31 , a determination unit (for example, an eye image determination processing unit) 32 that is an eye image determination processing unit, and an enlargement constituting an eye image output unit. Means (for example, target eye image enlargement processing unit) 33 and iris processing means (for example, iris processing unit) 34 are provided. The eye image determination processing unit 32 is connected to the output side of the image capture unit 31, and a desired eye image in which an eye is considered to be captured from one frame of the multiple divided images S <b> 31 b generated by the image capture unit 31. (This is referred to as a “target eye image”) and selected. The image capture unit 31 and the eye image determination processing unit 32 constitute a selection unit, and a target eye image enlargement processing unit 33 is connected to the output side of the eye image determination processing unit 32. The target eye image enlargement processing unit 33 enlarges the target eye image selected by the eye image determination processing unit 32 to the original size and outputs an eye image of one frame, and an iris processing unit 34 is provided on the output side. Is connected.

  An iris coding processing unit 34 a is provided in the iris processing unit 34. The iris coding processing unit 34a creates an iris code S34a from one frame of eye image sent from the target eye image enlargement processing unit 33, and an iris registration processing unit 34b is connected to this output side. Yes. The iris registration processing unit 34b registers the iris code S34a in the iris dictionary 34c together with the personal information of the identified person 14 when the iris code S34a is registered. Further, an iris collation processing unit 34d is connected to the output side of the iris coding processing unit 34a. The iris matching processing unit 34d performs personal identification by matching the iris code S34a sent from the iris coding processing unit 34a with the iris code registered in the iris dictionary 34c at the time of matching. The identification result is output.

  As shown in FIG. 6A, the frame switcher 20 in FIG. 5 receives, for example, identification target person images S16-1 to S16-n sent from a plurality of eye image cameras 16-1 to 16-n. An input buffer 21 is provided, and an image size reduction means 22 is connected to the output side of the input buffer 21. The image size reduction means 22 reduces the image size by performing a thinning process or the like of each identified person image S16-1 to S16-n sent from the input buffer 21, and the reduced image is displayed on the output side. Combining means 23 is connected. The reduced image synthesizing unit 23 synthesizes the reduced images reduced by the image size reducing unit 22 to generate one frame of a plurality of divided images S23 composed of analog video signals.

  As shown in FIG. 6B, the image capture unit 31 in FIG. 5 includes, for example, an A / D conversion unit 31a that converts the multi-segment image S23 of one frame sent from the reduced image synthesis unit 23 into a digital signal. A buffer 31b is connected to the output side of the A / D conversion means 31a. The buffer 31b outputs one frame of a plurality of divided images S31b composed of digital video signals sent from the A / D conversion means 31a at a predetermined timing and sends them to the eye image determination processing unit 32 in FIG.

  As shown in FIG. 7, the eye image determination processing unit 32 in FIG. 5 includes, for example, a pupil circle detection processing unit 32a that detects the pupil circle of the eye, an iris circle detection processing unit 32b that detects the iris circle, and an eye determination process. It consists of part 32c. The eye determination processing unit 32c determines and selects a target eye image in which the iris of the eye is most likely captured based on the detection results of the detection processing units 32a and 32b.

(Identification method)
FIG. 8 is a diagram showing the iris of the eye 15 in the person to be identified 14 in FIG. In the eye 15, there is a dark circular pupil 15a that is dark, and there is a slightly bright substantially circular iris 15c around the pupil circle 15b. The centers of the iris circle 15d and the pupil circle 15b are substantially the same. Match.

  FIGS. 9A and 9B are views showing an example of one frame of a plurality of divided images S23 formed of analog video signals formed by the frame switcher 20 of FIG. 6, and FIG. The divided image and FIG. 5B are 16 divided images.

  The personal identification device of Reference Example 1 is provided with a plurality of eye image cameras 16-1 to 16-n, and does not impose an unreasonable posture on the person to be identified 14 having various heights. It is designed to capture images correctly. The personal identification method will be described below with reference to FIGS.

  As shown in FIG. 4, when the identified person 14 stands on the right side of the cold mirror 13, the identified person 14 is photographed by the plurality of eye image cameras 16-1 to 16-n in the photographing process. The identified person images S <b> 16-1 to S <b> 16-n of each frame composed of analog video signals photographed by the eye image cameras 16-1 to 16-n are sent to the frame switcher 20 for the divided image generation processing.

  As shown in FIG. 6A, in the frame switcher 20, the identification target person images S16-1 to S16-n sent from the eye image cameras 16-1 to 16-n are input to the input buffer 21, and the images are displayed. The size reduction means 22 performs a thinning process or the like to generate a reduced image. The generated reduced images are combined into one frame image by the reduced image synthesizing unit 23 as shown in FIG. 8, and a one-frame multiple divided image S23 composed of analog video signals is generated. It is sent to the image capture unit 31.

  As shown in FIG. 6B, in the image capture unit 31, the one-frame multiple divided image S23 sent from the reduced image synthesizing unit 23 is converted into a digital signal by the A / D conversion unit 31a, and the buffer 31b. Through the above, one frame of a plurality of divided images S31b composed of digital video signals is output. The multiple divided image S31b is sent to the eye image determination processing unit 32 for selection processing.

  As shown in FIG. 7, the eye image determination processing unit 32 determines and selects a target eye image that is most likely to reflect the iris 15 c of FIG. 8 from the sent divided image S31 b. That is, in order to determine whether or not each image in the plurality of divided images S31b is a target eye image, first, the pupil circle detection processing unit 32a detects a circle that seems to be a pupil circle 15b from the image. In the case of the pupil circle 15b, as shown in FIG. 8, the inner side is dark and the outer side is bright. Therefore, a circle having such a feature may be detected. The circle detection process assumes a circle with a certain center and radius in the search range in the image and examines the density difference of the assumed circle edge (here, the inside is dark and the outside is bright). This is done by detecting the circle with the highest density difference. When the pupil circle 15b is detected, the detection result is sent to the iris circle detection processing unit 32b. If no circle is detected in the search range, the detection result is sent to the iris circle detection processing unit 32b as pupil circle detection failure.

  In the iris circle detection processing unit 32b, as in the pupil circle detection processing unit 32a, the iris circle 15d that is dark on the inside and bright on the outside is detected. When the iris circle 15d can be detected, the detection result is sent to the eye determination processing unit 32c. If no iris circle 15d is detected in the search range, this detection result is sent to the eye determination processing unit 32c as an iris circle detection failure.

  The eye determination processing unit 32c determines whether or not the eye 15 is in the image based on the detection results of the pupil circle detection processing unit 32a and the iris circle detection processing unit 32b. In this determination, whether or not the pupil circle 15b and the iris circle 15d have been successfully detected, and these circles 15b and 15d have substantially the same center, and whether the average image density of the area inside the pupil circle 15b is lower than a certain threshold value. And the target eye image is determined. In addition, when a plurality of candidates that the eye 15 appears to appear in each image in the multiple divided image S31b appear, the density difference between the edge portions of the pupil circle 15b and the iris circle 15d, and the positions of the two circles 15b and 15d The target eye image is selected by evaluating the image density of the pupil circle 15b region and selecting the image most likely to be the eye 15. The target eye image selected in this way is sent to the target eye image enlargement processing unit 33 for eye image output processing.

  In the target eye image enlargement processing unit 33, the target eye image that has been divided and reduced as shown in FIG. 9 is enlarged to the original size, and an eye image of one frame is output. This eye image is sent to the iris processing unit 34 for iris processing.

  In the iris processing unit 34, an iris code S34a is created from the received one-frame eye image in the iris coding processing unit 34a. At the time of registration, the iris code S34a is registered in the iris dictionary 34c together with the personal information of the identified person 14 in the iris registration processing unit 34b. On the other hand, at the time of collation, the iris collation processing unit 34d performs collation processing by comparing the generated iris code S34a with the iris code registered in the iris dictionary 34c, and outputs a personal identification result.

(effect)
In the personal identification device of Reference Example 1, the identified person 14 is photographed by a plurality of eye image cameras 16-1 to 16-n, and the photographed plural-frame identified person images S16-1 to S16-n are framed. Since the switcher 20 collectively generates one frame of a plurality of divided images S23, the host computer 30 selects a target eye image from the plurality of divided images S23 and performs registration / collation processing of the iris 15c. It is possible to identify an individual by photographing an eye image without forcing an unreasonable posture with respect to the person 14 to be identified having a different ID. In addition, a plurality of frames of the identified person images S <b> 16-1 to S <b> 16-n are collected by the frame switcher 20 to generate one frame of a plurality of divided images S <b> 23, and the plurality of divided images S <b> 23 are analyzed by the host computer 30. Therefore, the configuration of the eye image cameras 16-1 to 16-n is simplified, and the cost of the personal identification device can be reduced.

(Reference Example 2)
(Constitution)
FIG. 10 shows a reference example 2 of the present invention, and is a configuration diagram of the eye image determination processing unit 32 of FIG. 5, and is common to the elements common to the elements in FIGS. 5 and 7 showing the reference example 1. The code | symbol is attached | subjected.

  As shown in FIG. 7, the eye image determination processing unit 32 of Reference Example 1 includes a pupil circle detection processing unit 32a, an iris circle detection processing unit 32b, and an eye determination processing unit 32c. On the other hand, as shown in FIG. 10, the eye image determination processing unit 32 according to the second reference example is different from the pupil circle detection processing unit 32a and the iris circle detection processing unit 32b similar to FIG. It is comprised with the eye determination process part 32d. The fractal eye determination processing unit 32d determines and selects the target eye image by examining the fractal dimension of the iris region based on the processing results of the pupil circle detection processing unit 32a and the iris circle detection processing unit 32b.

(Identification method)
FIGS. 11A, 11B, and 11C are explanatory diagrams of the fractal eye determination processing unit 32d in FIG. 10, in which FIG. 11A is the original iris image, and FIG. 11B is the preprocessing. The iris image and the figure (c) are diagrams showing the iris image after the binarization process. Hereinafter, the personal identification method of the reference example 2 in the personal identification device of FIG. 5 will be described with reference to FIG.

  In the personal identification device shown in FIG. 5, as in the first reference example, the identified person 14 is photographed by a plurality of eye image cameras 16-1 to 16-n, and a frame-divided image S <b> 23 from the frame switcher 20. This is captured by the image capture unit 31 and sent to the eye image determination processing unit 32 as one frame of a plurality of divided images S31b made up of digital video signals. The eye image determination processing unit 32 determines and selects a target eye image that is most likely to reflect the iris 15c from a plurality of divided images S31b of one frame by the method shown in FIG.

  That is, in the eye image determination processing unit 32 shown in FIG. 10, the pupil circle 15b is detected by the pupil circle detection processing unit 32a, and then the iris circle 15d is detected by the iris circle detection processing unit 32b, as in FIG. The Thereafter, the fractal eye determination processing unit 32d determines whether or not the eye 15 is in the image based on the processing results of the pupil circle detection processing unit 32a and the iris circle detection processing unit 32b. That is, in the fractal eye determination processing unit 32d, when the detection of the pupil circle 15b and the iris circle 15d is successful, the target eye image is determined and selected by examining the fractal dimension in the region of the iris 15c as follows. Is done.

  When performing fractal analysis, it is necessary to convert to a binary image by binarization in order to calculate a fractal dimension for the gray scale image of FIG. If binarization is performed, a large amount of information in the image may be lost. Therefore, it is necessary to perform binarization while leaving the characteristic pattern of the original image as shown in FIG. 11A as much as possible. There is. For the iris image as shown in FIG. 11A, the gradation value of the image is non-uniform, and therefore, for example, adaptive binarization processing is performed. At this time, in the original iris image as shown in FIG. 11A, since the pattern of the iris 15c is difficult to identify, for example, in order to clarify the details of the image by distributing the gradation value of the image, binary Before smoothing, histogram smoothing is performed as preprocessing to obtain an image as shown in FIG. Thereafter, when an adaptive binarization process is performed, an iris image as shown in FIG. 11C is obtained.

  The fractal dimension is a quantity representing the complexity of a figure, and generally takes a non-integer value unlike empirical dimensions such as conventional points, straight lines, and planes. A figure on a two-dimensional plane takes a dimensional value from 0 to 2, and a more complicated figure has a larger value. Various methods for calculating the fractal dimension have been proposed, including a box count dimension, a Hausdorff dimension, a mass dimension, a Minkowski-Burigan dimension, a correlation dimension, and an information dimension. When the values of the iris region are examined, they take values in a substantially fixed range (for example, in the box count dimension in which how fine fractal structures are distributed on a plane is 1.6 ± 0. 05). Therefore, the target eye image is determined from the value 1.6 ± 0.05. In this case, the feature amount used in Reference Example 1 can also be used.

  The target eye image thus selected by the fractal eye determination processing unit 32d is sent to the target eye image enlargement processing unit 33 in FIG. In the target eye image enlargement processing unit 33, the target eye image is enlarged to the original size and sent to the iris processing unit 34. The iris processing unit 34 performs registration / collation processing in the same manner as in the first reference example.

(effect)
Reference Example 2 has the same effect as Reference Example 1. Furthermore, when the target eye image is determined and selected from the one-frame multiple-divided image S31b, the target eye image is determined using the fractal dimension. High target eye image can be determined.

(Reference Example 3)
(Constitution)
12 is an external view of a personal identification device showing Reference Example 3 of the present invention, FIG. 13 is a block diagram of FIG. 12, and FIG. 14 is a block diagram of the frame switcher 20A of FIG. Elements common to the elements in FIGS. 5 and 6 are denoted by common reference numerals.

  As shown in FIG. 12, in the personal identification device of Reference Example 3, a frame switcher 20A having a configuration different from that of the frame switcher 20 of FIG. 4 is provided, and a frame switcher control unit 35 is provided instead of the host computer 30 of FIG. The added host computer 30A is provided on the output side of the frame switcher 20A. The frame switcher control unit 35 controls the frame switcher 20A to output an output of a plurality of divided images S23 of one frame made up of analog video signals and to-be-identified person images S16-1 to S16-n for each frame made up of analog video signals. Has a function of switching between the output and the output.

  As shown in FIG. 13, in the host computer 30A, an image capture unit 31, an eye image determination processing unit 32, and an iris processing unit 34 similar to those in FIG. 5 are provided, and a target eye image enlargement processing unit 33 in FIG. Instead, an image acquisition control unit (for example, a frame switcher control unit) 35 and a recapture unit (for example, an eye image recapture unit) 36 are provided. The frame switcher control unit 35 and the eye image recapture unit 36 constitute an eye image output unit. Among these, the eye image recapture unit 36 is one frame composed of an analog video signal sent from the frame switcher 20A. The identified person image (any one of S16-1 to S16-n) is subjected to A / D conversion to output a one-frame eye image composed of a digital video signal and send it to the iris processing unit 34.

  As shown in FIG. 14, the frame switcher 20A is configured by adding a switching means 24 to the frame switcher 20 of FIG. The switching unit 24 is connected, for example, between the output side of the input buffer 21 and the input side of the image size reduction unit 22, and from an analog video signal sent from the input buffer 21 under the control of the frame switcher control unit 35. The identified person images S16-1 to S16-n for each frame are output to the image size reduction means 22 as they are, or any one of the identified person images S16-1 to S16-n is subjected to an iris process. Whether to output to the unit 34 is switched.

(Identification method)
In the identification method using the personal identification device of Reference Example 3, as shown in FIG. 13, the eye image output processing in the frame switcher control unit 35 and the eye image recapture unit 36 is the target eye image enlargement processing unit in FIG. This is different from the eye image output processing at 33. The identification method will be described below.

  In FIG. 13, the divided image generation processing by the frame switcher 20A and the image capture unit 31 and the selection processing by the eye image determination processing unit 32 are performed in the same manner as the identification method of FIG. That is, from the analog video signal by the frame switcher 20A from the identified person images S16-1 to S16-n for each frame made up of analog video signals photographed by a plurality of eye image cameras 16-1 to 16-n. A one-frame multiple-divided image S23 is generated. The multiple-divided image S23 is captured by the image capture unit 31 and converted into a single-frame multiple-divided image S31b composed of a digital video signal. The eye image determination processing unit 32 determines the target eye image from the multiple-divided image S23. Selected.

  In FIG. 5 showing the reference example 1, after the selection process by the eye image determination processing unit 32, the target eye image is enlarged from the multiple divided image S31b by the eye image output process of the target eye image enlargement processing unit 33, so that a normal eye image is output. The image is enlarged to the image size and sent to the iris processing unit 34. However, when the number of eye image cameras 16-1 to 16-n increases, the image in the multi-divided image S31b has to be reduced. In this case, if the target eye image enlargement processing unit 33 performs the enlargement process from the multiple divided image S31b as shown in FIG. 5, the information of the original eye image may be excessively dropped.

  Therefore, in FIG. 13 showing the reference example 3, when the target eye image is determined by the eye image determination processing unit 32, the frame switcher control unit 35 recognizes the plurality of eye image cameras 16-1 to 16-n so far. The person images S16-1 to S16-n output from the frame switcher 20A as the multiple divided images S23, the frame switcher 20A is controlled to switch the output side of the switching means 24 to the eye image recapture unit 36 side. . Then, an eye image is captured again by the eye image camera (any one of 16-1 to 16-n) where the target eye image is captured, and this eye image is output from the switching means 24 in the frame switcher 20A. . This eye image is captured by the eye image recapture unit 36 and converted into a one-frame eye image composed of a digital video signal. This makes it possible to acquire a target eye image without performing the enlargement process as shown in FIG. This eye image is sent to the iris processing unit 34, and registration / collation processing is performed as in FIG.

(effect)
Reference Example 3 has substantially the same effect as Reference Example 1. However, when the target eye image is determined by the eye image determination processing unit 32, instead of enlarging the eye image from the plurality of divided images S31b as in Reference Example 1, the eye image is re-taken again, and the eye image re-capturing unit 36 Therefore, a more accurate eye image can be obtained, and more accurate registration / collation processing can be performed.

Example 1
(Constitution)
1 is an external view of a personal identification device showing Embodiment 1 of the present invention, and FIG. 15 is a block diagram of FIG. 1. Elements common to those in FIGS. 4 and 5 showing Reference Example 1 are common. The code | symbol is attached | subjected.

  As shown in FIG. 1, in the personal identification device according to the first embodiment, sensors (for example, proximity sensors) 18-1 to 18-n are provided beside the eye image cameras 16-1 to 16-n in FIG. 4. is set up. Each proximity sensor 18-1 to 18-n detects the distance to the person to be identified 14, and this output side is sensor control provided in a host computer 30B having a configuration different from that of the host computer 30 in FIG. Connected to the unit 37. The sensor control unit 37 in the host computer 30B acquires the detection results of the proximity sensors 18-1 to 18-n.

As shown in FIG. 15, the host computer 30B is configured by adding a sensor control unit 37 and an unnecessary eye image detection processing unit 38 to the host computer 30 of FIG. The sensor control unit 37 and the unnecessary eye image detection processing unit 38 constitute an image selection unit, and among these, the unnecessary eye image detection processing unit 38 is an image capture unit 31 and an eye image determination processing unit 32 which is an eye image determination processing unit. The height of the identified person 14 is estimated from the detection results of the proximity sensors 18-1 to 18-n acquired by the sensor control unit 37, and one frame of the image captured by the image capture unit 31 is converted. An eye image that is considered to have an eye is roughly selected from the plurality of divided images S31b, and the selection result is sent to the eye image determination processing unit 32.

(Identification method)
FIG. 16 is a diagram illustrating how the height is detected by the proximity sensors 18-1 to 18-n in FIG. Hereinafter, a personal identification method using the personal identification device of FIGS. 1 and 15 will be described with reference to FIG.

  In the personal identification device of FIG. 15, as in the first reference example, identification target person images S <b> 16-1 to S <b> 16 -n photographed by a plurality of eye image cameras 16-1 to 16-n include the frame switcher 20 and the images. By the divided image generation processing by the capture unit 31, a one-frame multiple divided image S31b composed of a digital video signal is generated. In Reference Example 1, eye image determination processing is performed using the multi-divided image S31b. On the other hand, in the present Example 1, the distance to the identified person 14 is detected by the proximity sensors 18-1 to 18-n installed beside the respective eye image cameras 16-1 to 16-n. The detection result is acquired by the sensor control unit 37 and sent to the unnecessary eye image detection processing unit 38.

  In the unnecessary eye image detection processing unit 38, the uppermost proximity sensor 18-3 that detects the proximity of the identified person 14 based on the detection result sent from the sensor control unit 37 is the top of the identified person 14. Presence is estimated (that is, the height of the person to be identified 14 is estimated), the proximity sensors 18-3, 18-4, and 18-5 that the eye 15 is supposed to be captured are predicted, and the eye 15 is clearly visible Except for the image that seems not to be present, an image that is likely to contain the eye 15 is roughly selected. From this selection result, as in Reference Example 1, the eye image determination processing unit 32 determines and selects a desired eye image in which the eye 15 appears to be captured, and the selected target eye image is the target eye image. The image is enlarged by the enlargement processing unit 33. The enlarged one-frame target eye image is subjected to registration / collation processing by the iris processing unit 34 as in the first reference example.

(effect)
In the first embodiment, since the proximity sensors 18-1 to 18-n, the sensor control unit 37, and the unnecessary eye image detection processing unit 38 are provided, the eye image determination processing unit 32 narrows down the eye image to be examined. The eye image to be determined when selecting is reduced. As a result, the target eye image can be determined at high speed and with high accuracy, and the entire registration / collation processing can be performed at high speed and with high accuracy. Furthermore, substantially the same effect as in Reference Example 1 can be obtained.

(Usage form, etc.)
The present invention is not limited to the above-described reference examples and examples, and various usage forms and modifications are possible. For example, the following usage forms and modifications are as follows.

(A) Reference Examples 1, 2, 3 and Example 1 can be configured in combination. For example, the eye image determination processing unit 32 in Reference Example 3 and Example 1 can be configured in FIG. 10 of Reference Example 2 in addition to the configuration in FIG. Further, the frame switcher control unit 35 and the eye image recapture unit 36 of FIG. 13 showing the reference example 3 can be used instead of the target eye image enlargement processing unit 33 of FIG. 15 showing the first embodiment.

(B) The frame switchers 20 and 20A in the reference examples 1, 2, and 3 and the first embodiment can be changed to, for example, a sequential switcher. The sequential switcher has one image input, but by controlling connection to the plurality of eye image cameras 16-1 to 16-n, the plurality of eye image cameras 16-1 to 16- Since it is possible to acquire images from n, if these plural images are output as a plurality of divided images, the same function as that of the frame switchers 20 and 20A can be achieved.

(C) In Reference Examples 1, 2, 3 and Example 1, a plurality of eye image cameras 16-1 to 16-n are arranged in one row in the vertical direction, but these are arranged in a plurality of rows in the vertical direction. May be. Accordingly, it is possible to cope with the case where the person to be identified 14 stands out of position, and the target eye image can be accurately acquired from the person 14 to be identified.

BRIEF DESCRIPTION OF THE DRAWINGS It is an external view of the personal identification device which shows Example 1 of this invention. It is an external view of the conventional personal identification device. FIG. 3 is a configuration diagram of FIG. 2. It is an external view of the personal identification device which shows the reference example 1 of this invention. It is a block diagram of FIG. It is a block diagram of the frame switcher 20 and the image capture part 31 of FIG. It is a block diagram of the eye image determination process part 32 of FIG. It is a figure which shows the iris of the eye 15 of FIG. It is a figure which shows the example of the several division image of FIG. FIG. 6 is a configuration diagram of an eye image determination processing unit 32 in FIG. 5, showing Reference Example 2 of the present invention. It is explanatory drawing of the fractal eye determination process part 32d of FIG. It is an external view of the personal identification device which shows the reference example 3 of this invention. It is a block diagram of FIG. It is a block diagram of the frame switcher 20A of FIG. It is a block diagram of FIG. It is a figure which shows a mode that a height is detected by the proximity sensor 18-1 to 18-n of FIG.

Explanation of symbols

13 cold mirror 14 identified person 15 eyes 15a pupil 15c iris 16-1 to 16-n eye image camera 18-1 to 18-n proximity sensor 20, 20A frame switcher 30, 30A, 30B host computer 31 image capture unit 32 eyes Image determination processing unit 32a Pupil circle detection processing unit 32b Iris circle detection processing unit 32c Eye determination processing unit 32d Fractal eye determination processing unit 33 Target eye image enlargement processing unit 34 Iris processing unit 35 Frame switcher control unit 36 Eye image recapture unit 37 Sensor control unit 38 Unnecessary eye image detection processing unit

Claims (2)

A plurality of eye image cameras that are arranged in one or more rows in the vertical direction, which is the height direction of the person to be identified, and that respectively photograph the person to be identified and output a plurality of images of the person to be identified;
A plurality of sensors installed beside each of the plurality of eye image cameras and detecting a distance to the identified person ;
Eye image determination processing means for determining and selecting an eye image from which the eye iris is most likely to be captured from the plurality of identification target images ,
Wherein the detection result of the sensor to estimate the height of an identified person, from the plurality of the plurality taken respectively eye image camera the identified person image, rough the eye images the eye is likely to show From the selection result, the eye image determination processing means determines and selects the eye image that the eye is supposed to be in, and performs personal identification using the iris pattern in the eye image. A personal identification device characterized by having a configuration. A one-way transmissive cold mirror disposed in the longitudinal direction, which is the height direction of the identified person,
One or a plurality of rows are arranged in the vertical direction, which is the height direction of the identified person, on the opposite side of the cold mirror facing the identified person, and a plurality of identified person images are obtained by photographing each identified person. A plurality of eye image cameras to output,
A plurality of sensors installed beside each of the plurality of eye image cameras and detecting a distance to the identified person ;
Eye image determination processing means for determining and selecting an eye image from which the eye iris is most likely to be captured from the plurality of identification target images ,
Wherein the detection result of the sensor to estimate the height of an identified person, from the plurality of the plurality taken respectively eye image camera the identified person image, rough the eye images the eye is likely to show From the selection result, the eye image determination processing means determines and selects the eye image that the eye is supposed to be in, and performs personal identification using the iris pattern in the eye image. A personal identification device characterized by having a configuration.

Images