JPS62272378A - Fingerprint collating device - Google Patents

Abstract

PURPOSE:To obtain stable directional angle of rising lines to improve the precision of collation by dividing the processing area of a fingerprint picture into small areas in accordance with the pitch width of rising lines of the fingerprint. CONSTITUTION:It is checked by a contrast check means 1-1 whether a fingerprint picture is fetched a frame buffer 4 or not, and the fingerprint picture is fetched from the frame buffer 4 by a picture input means 1-2. The pitch width of rising lines of the fingerprint is measured by a pitch width measuring means 1-3, and the processing area is divided into small areas, whose size is 10X10 pixels, by a small area dividing means 1-4 if the pitch width is equal to or longer than 5 pixels, and it is divided into small areas whose size is 8X8 pixels if the pitch width is shorter than 5 pixels. Directional angles of rising lines in small divided areas are calculated by a directional angle calculating means 1-5, and directional angle data are registered by a register means 1-6. Collation data and register data are collated with each other by a collating means 1-7 to discriminate whether they coincide with each other or not.

Description

[Detailed Description of the Invention] 3. Detailed Description of the Invention [Field of Industrial Application] The present invention relates to a fingerprint verification device that performs fingerprint verification to authenticate individuals entering and exiting a specific controlled area.

[Conventional technology]

Figure 5 is a schematic configuration diagram of a conventional fingerprint verification device.
1) is a prism, (2) is a finger rest installed on the prism (1), (3) is a camera that takes an image of the fingerprint sent through the prism, (0 is a camera that takes an image of the fingerprint sent through the prism (3), (5) is a light source that illuminates the area where the finger is placed when the fingerprint image is captured by the camera (3); (6) is the numeric keypad for entering the password;
7) is the mode changeover switch, (8) is the relay, (9
) is a door connected to relay (8), and when it is determined that the door belongs to the same person as a result of verification, relay (8) turns OFF.
Press N to open the door (9).

In the 5th rA, illuminate the prism (1) with the light source (5), operate the numeric keys (6) and enter the password, and the camera (3) will be able to capture the image. When the finger is pushed out onto the stand (2), the finger is captured by the camera (3) and the image data of the finger is stored in the frame buffer (4).

When the mode switch (7) is in verification mode, the stored registered data corresponding to the entered PIN number is verified against the newly input verification data via the frame buffer (4), and both are the same person. When it is determined that the object is the object, the relay (8) is turned on and the door (9) is opened.

When the mode switch (7) is in the registration mode, the password and image data are registered. For example, the image data at the time of verification or registration is n
A method is used to find the direction angle of each ridge for each small region.

[Problem that the invention seeks to solve]

Since the conventional fingerprint verification device is configured as described above, it is necessary to obtain stable ridge direction angle data and register or verify the direction angle data. However, because there is a difference in pitch width of the ridges due to bias, if the number of divisions of the processing area is fixed, the number of ridges that will fit in a small area will differ, and if multiple ridges are included, the average direction There was a problem in that the angle data had to be obtained, and the direction angle data was obtained in an unstable form.

This invention was made to solve the above-mentioned problems, and it is possible to change the size of the small area depending on the pitch width, obtain stable direction angle data, and furthermore, it is possible to obtain stable directional angle data. The purpose of the present invention is to obtain a one-finger mosquito verification device that can improve verification accuracy by registering the pinch width and comparing the pitch widths during verification.

[Means for solving problems]

The fingerprint verification device according to the present invention includes a pitch width measurement device that measures a pinch width of a ridge of a fingerprint from a fingerprint image of a person to be verified that is captured by an image input device, and a processing area of the fingerprint image based on the measured pitch width. The apparatus includes a small area dividing means for dividing into small areas, and a directional angle calculation means for calculating directional angle data of ridges in the small area to generate fingerprint verification data.

[Effect]

In this invention, when the pitch width measuring means measures the pinch width of the ridges of the fingerprint which are different from each other due to bias from the fingerprint image, the small area dividing means divides the processing area of the fingerprint image into small areas according to the measurement result. Since the area is divided into two, multiple ridges do not enter each small area. Therefore, the direction angle calculation means can obtain stable ridge direction angle data from each small region.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. 1st
In the figure, (+-1) indicates that the fingerprint image is at frame bump 7.
(4) contrast check means for determining whether the image has been captured; (1-2) is image input means for capturing the image; (1-3) is pitch width measuring means for measuring the pitch width of the ridge; -4) is a small area dividing means for dividing the image processing area into several small areas, (+-5) is a direction angle calculation means for calculating the direction angle of the ridge that enters each small area, and (16) is the direction A registration means (1-7) is used to register corner data, and a collation means (1-7) is used to compare newly captured image data with registered image data.

Next, an embodiment of the present invention will be described based on the above configuration. The contrast check means (1-1) checks whether or not the fingerprint image has been captured in the frame buffer (4), and if it has been captured, the image input means (1-2) inputs the fingerprint image from the frame buffer (4). Measure the pitch width of the ridges of the fingerprint using a pitch width measuring means (+-3) from the fingerprint image captured in this way, and if the pinch width is 5 pixels or more, the size of the small area is determined. I
OX 10 pixels, and 5 pixels if it is a non-cat, then 8
The processing area is divided into ×8 pixels by a small area dividing means.

The direction angle of the ridge within the divided small area is determined by the direction angle calculation means (
+-5) and this direction angle data is registered by the registration means (IJ). The direction angle data, that is, the verification data obtained in this way is compared with the registered direction angle data, that is, the registered data, by the verification means (1-7), and it is determined whether the data are the same.

FIG. 2 shows the hardware configuration for implementing each of the above means. The hardware configuration includes a finger rest (2) on which to place a finger, a light source (5), a prism (1) to apply light, and a camera (3) to project an image. ), an A/D converter (2a) that digitizes the image,
It consists of a frame buffer (4) for storing digitized image data, a microprocessor (2b), and a registration data buffer (2c).

Next, the operation of this embodiment according to the above configuration will be explained with reference to FIGS. 3 and 4. FIG. 3 is a flowchart explaining the operation of this embodiment, and FIG. 4 is a detailed explanatory diagram for calculating the pitch width of the ridge.

First, when performing fingerprint verification, check whether it is registration mode or verification mode in (101). If it is registration mode, go to (102), and if verification mode, go to (120).
). At (102), a guidance to enter the PIN is displayed, and when the PIN is entered, at (103), this PIN is imported, and at (104), the 4
It is checked whether the digit PIN number is already registered, and if it is registered, the process returns to (102). If it is not registered, proceed to (105). In (105), the contents of the 2-frame buffer (4) with no finger placed are taken in. At this time, the density variance of each 16×1B pixel square at each of the four corners of a predetermined processing area (for example, 120 pixels×80 pixels) is obtained and stored in QWrNI BUN. (10ft) displays finger placement guidance, (+07) performs the same processing as (+05) and stores the variance obtained in QW BUN, (10B) displays the variance of the four corners obtained at (+05),
Find the respective ratios of the variances of the four corners found in (107),
If the smallest ratio among the four ratios is 3.5 or more, it is determined that a finger has been placed, and the process proceeds to (109). If the ratio of the variances of the four corners is smaller than 3.5, for example, start over from (10B).

At (+09), the processing area is binarized with O°” and 1゛,
Consider a small region of 8×8 pixels in the center, and find the direction angle of the ridge that falls within this small region. (110) then (109
) Count the number of pixels for 4 pitches in the perpendicular direction to the direction angle found in . In (111), the average of these four pitches is taken and stored in KQPCH, and this value is used as the pitch width of the fingerprint image currently being processed.

(112) checks whether this pitch width KQPCH is, for example, 5 pixels or more, and if it is 5 or more, (114)
Then, the processing area is further divided into several 12×8 pixel square small areas and the division flag is set as KQPC)IF-1. When KQPCH is less than 5 pixels (113)
Divide the processing area into several 15x10 pixel square small areas and use the division flag KQPC)! Let it be F-0.

In (115) and (11B), the direction angle of the ridge in each small region is determined, and if the total number of direction angles found is, for example, 80% or more of the total number of small regions, go to (117) and perform the calculation in (103). Entered PIN, image direction data, and division flag K
QPC) Register IF L4 (+01) Return to (11
B) If it is less than 1llO%, return to (toe) and retake the image.

(102) Kara (117) Te is the registration process (
In step 120), the PIN number and image are newly imported, direction angle data and division flag KQPCHF are obtained, and direction angle data and division flag KQPC, which are registered as a pair with this PIN number, are obtained.
) IF 't, , :KoteKQPcHF,=
! KQPC) IF 'te, the pinch width is different, so it is determined that the image is not a field image. If KQPCHF=K
QPCHF”, calculate the average of the squares of the differences between the two sets of direction angle data, use this as the distance between the two sets of direction angle data, and set this distance to a value smaller than a predetermined threshold (for example, 15 degrees squared). If so, it is determined that they are the same image, and if it is equal to or greater than the threshold value, it is determined that the images are different and the process returns to 101).If the images are the same, the door (3) is opened for one second and the process returns to 101).

The detailed operation of step (+10) above will be explained with reference to FIG. Direction angle determined by (+09) (Direction 1 in the diagram)
Consider the direction perpendicular to (direction 2 in the figure). Find the angle θ between direction 2 and the X-axis. Here 1:s:nθ-p:
Find the least common multiple p and q of q. Considering the positive direction of X from the center, repeat p pixels in the X direction and -q pixels in the y direction, so 0 to 1 or 1 to 0
Find the first point that changes to and set its coordinates to (z, y). Roughly moving P pixels in the X direction and -q pixels in the Y direction, count the number of pixels for 2 pitches from (=, y). Next, consider the negative direction of: from the point (x, y): -p pixels in the x direction and q pixels in the y direction, and attack the number of pixels equivalent to 2 pixels.

The present invention is not limited to the above embodiments, and the following embodiments are also possible.

(1) Divide the processing area into 15x 10 pixels and 12x
Although two types of 8 pixels are used, three or more types may be used.

(2) Use CRT, liquid crystal, etc., or install two lamps for PIN entry guidance and finger placement guidance display.
Guidance may be provided by lighting up each of these lamps.

(3) When calculating the pitch width, it is determined by averaging 4 pitches, but instead of measuring at 5 pitches or more or at one location, it may be measured at random positions and averaged.

〔Effect of the invention〕

As described above, according to the present invention, the processing area is divided into two ways depending on the pitch width of the ridge, and the small area in which the direction angle is calculated can be changed depending on the thickness of the ridge, so the direction angle is more stable. It is possible to perform verification at a stable direction angle during verification, and adding the size of the small area to the registered data has the effect of further increasing the verification rate and obtaining high verification accuracy.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram of an embodiment of the fingerprint verification system of the present invention, and FIG. 2 is a hardware configuration diagram configuring the above embodiment.
FIG. 3 is a flowchart for explaining the operation of the above embodiment, FIG. 4 is a detailed explanatory diagram of the pitch width calculation routine in the above flowchart, and FIG. 5 is a hardware configuration diagram of a conventional fingerprint collation device. In the figure, (2) is a finger rest, (3) is a camera, (4) is a frame buffer, (1-2) is an image input means, (1-3) is a pitch width measurement means, (+-4) is a small area dividing means, (1-5) is a direction angle calculation means, (16) is a registration means, (
1-7) is a verification means. In addition, in the figures, the same reference numerals indicate the same or equivalent parts. Agent Masuo Oiwa Figure 1, Figure 2, Figure 4 - Written amendment to the original procedure (voluntary) 1925/2015 Commissioner of the Japan Patent Office y I & 61 10 21 1. Indication of case Japanese Patent Application No. 116932/1983 Week 2, Name of invention Fingerprint verification device 3, person making the correction Representative Moriya Shiki 5, subject of correction ＼ζshi 6, content of correction (1) “Therefore, direction angle calculation” on page 5, lines 6 to 7 of the specification ``The means can obtain stable direction angle data of the ridge from each small region.'' has been changed to ``Therefore, the direction angle calculation means can more stably obtain the direction angle of the ridge in each small region.'' to correct. (2) "Identical image/
...Return to 101) for 1 second. ” is replaced with “
It is determined that they are the same image, the door (9) is opened for one second, and the process returns to step (101). If the images are equal to or greater than the threshold, it is determined that the images are different and the process returns to step (101). ” he corrected. (3) r I:s:nθ-P on page 10, line 11 of the specification
:(TJ to r sinθ−P :(T
Correct it with J. (4) The description "fingerprint verification total" on page 12, line 4 of the specification is amended to read "fingerprint verification device." (5) Correct the drawing medium temperature diagram 3 as shown in the attached sheet. 7. At least 1 copy of catalog drawings of attached documents

Claims (1)

[Claims] In a fingerprint verification device that captures personal verification data based on a fingerprint image using an image input means and determines whether the verification data and registered data belong to the same person, the pitch width of the ridges of the fingerprint is determined from the captured fingerprint image. A pitch width measuring means for measuring, a small area dividing means for dividing a processing area of a fingerprint image into small areas according to the measured pitch width of the ridges, and a direction for calculating the direction angle of the ridge entering each small area. an angle calculation means; a registration means for registering the calculated direction angle data;
1. A fingerprint verification device comprising: verification means for determining whether registered direction angle data and captured direction angle data for comparison belong to the same person.