JPH09198485A - Method and device for detecting rotary fingerprint - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a sharp rotary fingerprint by measuring a rotary fingerprint picture on an optical plane as an image, cutting out the image of a prescribed width from a center line orthogonal to the rotating direction of that image toward along the rotating direction into a strip shape and compositing these cut-out images. SOLUTION: The fingerprint picture inputted from a two-dimensional sensor is converted to multi-gradation digital data by an A/D converter 7. The converted data are sent to a center detection circuit 9. The center detection circuit 9 finds the position of center line of each sample fingerprint picture. The center detection circuit 9 sends the found coordinates of the center line to a write control circuit 11. Since certain processing time is required for the center detection circuit 9 and the write control circuit 11, the digital data from the A/D converter 7 are sent to a memory 10 for composition after time is controlled by a delay circuit 8. According to the command of the write control circuit 11, the memory 10 for composition composites the data. After this composition, the data are displayed through a D/A converter 12 onto a monitor 14 as the rotary fingerprint.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a rotary fingerprint detecting method and device for automatically collecting a rotary fingerprint obtained by rotating a finger.

[0002]

2. Description of the Related Art FIG. 3 shows an optical system block of a rotary fingerprint input device having an optical prism. Here, the rotating fingerprint is
The fingerprint around the finger obtained by rotating the finger.
Since the finger is round and the fingerprint around the finger cannot be obtained by merely pressing the finger once, it is necessary to detect the rotating fingerprint. The rotary fingerprint detecting device automatically detects the rotary fingerprint, and the rotary fingerprint input device performs input at that time. In this figure, the totally reflected light is made incident on the sensor, and when the finger is not placed on the optical prism 1, the light 6 emitted from the light source 4 is totally reflected by the optical surface 2 of the prism and is reflected by the two-dimensional sensor 5. The light source 4, the prism 1, and the two-dimensional sensor 5 are arranged so as to be incident. FIG. 4 shows an input method using total reflection light. When the finger 3 is placed on the optical surface 2 on the prism, the ridge portion of the fingerprint contacts the optical surface 2,
Since light is scattered at this portion, only a part of the light is incident on the sensor. This state is shown by light ray 6A. On the other hand, since the valley portion of the fingerprint does not contact the optical surface 2, the light ray 6B is totally reflected and almost all the light enters the sensor 5. In this way, the fingerprint can be input to the sensor 5 with good contrast. There is also a method of making scattered light enter the sensor. In this case, the total reflection light of the light source is arranged so as not to enter the sensor, and when the finger is placed on the prism, the sensor detects the light scattered by the ridge.

5 and 6 show a method of synthesizing a rotary fingerprint from a rotary fingerprint image detected by the sensor 5. In FIG. 5, the finger smoothly rotates on the optical surface 2. At a certain time T = t ₁ , T = t ₂ , T = t ₃ , the position of the finger 3
It shows with a, 3b, and 3c. Two-dimensional sensor 5 corresponding to this
Are respectively the fingerprint image of the finger 3a, the fingerprint image of the finger 3b, and the fingerprint image of the finger 3c. Such time-continuous fingerprint images are combined to obtain a rotated fingerprint combined image as shown in FIG.
Note that the positions of the fingers 3a, 3b, and 3c in FIG. 5 are separated from each other for the sake of explanation, but they are actually closer to each other.

As a method for obtaining a rotated image by rotating a finger on such an optical surface, there is disclosed a "fingerprint image input device" in JP-A-2-171881, a "fingerprint identification device" in JP-A-5-101168, and JP-A-6- 500654 "Method and apparatus for capturing skin trace images". JP-A-2-1
In the “Fingerprint image input device” of 71881, the area of the rotated image is divided into strips in advance, it is determined that the fingerprints have been input to the area, and the strips corresponding to the input fingerprints are cut out and combined. Is the way to go. Japanese Patent Application Laid-Open No. 5-101168 describes that a "fingerprint identifying device" synthesizes vertical columns around a center line orthogonal to the center or rotation direction of a fingerprint image input at a certain time. There is no description about the method. JP 6
In “-500654“ Method and apparatus for complementing skin trace image ”, pixel-by-pixel replacement is performed according to image intensity. That is, if the intensity of brightness is higher than that of the fingerprint image that has already been input, it is a method of replacing each pixel.

[0005]

In contrast to the above prior art,
Further, an experiment was conducted to take measures against the dynamic distortion caused by the rotation of the finger. As a result of the experiment, a certain time t _i
It was found that there was an unclear area when the fingerprint image of was analyzed. Each area of the fingerprint image will be described with reference to FIG.
The region E ₁ is a region where the finger is approaching, and the region that has been pressed so far is unclear and has a large distortion. The area E ₂ is just the pressed area and the sharp area. The area E ₃ is a portion to be pressed from now on and is an unclear area. A center line C orthogonal to the rotation direction of the fingerprint image (the center line is the area E in FIG.
A straight line that passes through the image centroid position in the fingerprint image consisting of ₁ , E ₂ , and E ₃ and that is orthogonal to the rotation direction. Alternatively, a straight line passing through the center positions of the left and right ends of the fingerprint image, for example, the left and right end positions are coordinate values in the rotation direction of a ₁ ,
If it is a ₂ , the center line is a straight line passing through the position of a ₁ + (a ₁ + a ₂ ) / 2. ), The area E ₂ is displaced from the center line C in the rotational direction. That is, the center line C extends to the right side rather than the left side. Therefore, when synthesizing the fingerprint images, it is necessary to consider the shift in the rotation direction.

On the other hand, when the finger is rotated, distortion (dynamic distortion) may occur in the rotation itself, but no solution has been proposed in the conventional example. Dynamic distortion due to the rotation of the finger will be described with reference to FIGS. The Figure 8 and 9, cut at d ₁ to d ₂ become arbitrary section for the image G _1, performs a cut in d ₂ to d ₃ becomes arbitrary section to the image G _2, synthesized by arranging them It is an example. If d ₁ and d ₂ are the center lines of the images G ₁ and G ₂ , this combination is related to the second embodiment described later, and provides the subject. FIG. 8 shows a case where the fingerprint is displaced during rotation, in a direction orthogonal to the rotation direction. In FIG.
For a fingerprint image G ₁ sampled at a certain time T = t _i , T =
At t _{i + 1} , a fingerprint image G ₂ in which the finger is rotated and is displaced in the long axis direction at the same time is shown. When the two are combined, as shown on the right side of FIG.
Then, it becomes discontinuous like the ridge 10b. FIG. 9 shows a composite image when the finger is displaced around an axis orthogonal to the optical surface during rotation. In this case, as shown on the right side of the drawing, the ridges 10a and 10b, which originally did not branch, are erroneously combined in the combined image.

It is an object of the present invention to provide a rotary fingerprint detecting device which enables clear detection of a rotary fingerprint by combining images in consideration of the rotation direction. A further object of the present invention is to provide a rotary fingerprint detecting device capable of eliminating erroneous composition due to rotational distortion. A further object of the present invention is to provide a rotary fingerprint detection device that enables correction of rotational distortion.

[0008]

SUMMARY OF THE INVENTION The present invention provides a rotary fingerprint detecting method in which a finger is rotated on an optical surface, and the whole fingerprint image is detected by measuring the finger. For each image thus obtained, an image of a predetermined width facing the rotation direction is cut into strips from the center line orthogonal to the rotation direction of the image, and the cut images are combined to obtain the entire image of the rotated fingerprint. Disclosed is a rotating fingerprint detection method.

Furthermore, the present invention provides a method for detecting a rotary fingerprint, in which a finger is rotated on an optical surface, and this is measured to detect an entire image of a rotary fingerprint, the rotary fingerprint image on the optical surface is measured as an image, and thus obtained. For each image, cut the image into strips in the section from the centerline position orthogonal to the rotation direction of the image to the centerline position of the next image along the rotation direction, combine the cutout images, and rotate. Disclosed is a rotating fingerprint detecting method for obtaining an entire fingerprint image.

Further, the present invention provides a method for detecting a rotary fingerprint, in which a finger is rotated on an optical surface and is measured to detect an entire image of a rotary fingerprint, the rotary fingerprint image on the optical surface is measured as an image, and thus obtained. For each image, an image with the maximum sample width (W _m ) is cut into strips from the center line orthogonal to the rotation direction of the image, and the cut images are combined to obtain a rotated fingerprint whole image. A fingerprint detection method is disclosed.

Further, the present invention discloses a rotary fingerprint detecting method for detecting the presence or absence of the rotational distortion of the finger at the composition boundary and correcting the rotational distortion if it occurs.

Further, the present invention discloses a rotating fingerprint detecting method in which when the cutout section is larger than a predetermined first reference value (W ₀ ), the cutout is notified as a defect.

Further, according to the present invention, the maximum sample width W _m is
A rotating fingerprint detecting method is disclosed in which the cutout is notified as a defect when the centerline position of the next adjacent image is not reached.

Further, the present invention provides a means for picking up a rotating fingerprint image on an optical surface in a rotating fingerprint detecting device for rotating a finger on the optical surface and measuring the rotation to detect an entire image of the rotating fingerprint. For each screen, from the center line orthogonal to the rotation direction of the image, a means for cutting out an image of a predetermined width in a strip shape facing the rotation direction, and means for synthesizing the cut-out images and detecting the entire image of the rotating fingerprint. , A rotary fingerprint detecting device comprising:

Further, the present invention provides a means for picking up a rotating fingerprint image on an optical surface in a rotating fingerprint detecting device for rotating a finger on the optical surface and measuring the rotation to detect the entire image of the rotating fingerprint, and thus obtained. For each screen, from the centerline orthogonal to the rotation direction of the image, the section from the centerline position of the next image along the rotation direction, means for cutting out the image in strips,
Disclosed is a rotary fingerprint detecting device including a unit for synthesizing the cut-out images to detect the entire image of the rotary fingerprint.

Further, the present invention provides a means for picking up a rotating fingerprint image on an optical surface in a rotating fingerprint detecting device for rotating a finger on the optical surface and measuring this to detect an entire image of a rotating fingerprint, and thus obtained. For each screen, a means for cutting out an image of the maximum sample width (W _m ) in the direction of rotation from the center line orthogonal to the direction of rotation of the image into strips, and the cut-out images are combined to produce the entire image of the rotated fingerprint. And a rotating fingerprint detecting device comprising a means for detecting the.

Further, the present invention discloses a rotary fingerprint detecting apparatus which is provided with means for detecting the presence or absence of the rotational distortion of the finger at the synthetic boundary and correcting the rotational distortion if the rotational distortion occurs.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described. As shown in FIG. 7, the clear area E ₂ has a characteristic of expanding in the rotation direction. Therefore, when the fingerprint image is cut into strips and synthesized as a rotated fingerprint image, strips of each fingerprint image are cut out with reference to the center line C, and the fingerprint images existing in the rotation direction from the center line C are compared. Take the method of cutting out a part. A rotated fingerprint image is obtained by synthesizing (connecting) strip-shaped images cut out for each fingerprint image.

Next, a second embodiment will be described. In the first embodiment, the region existing in the rotation direction with respect to the centerline C is cut out, but the cutout width is not particularly limited. It is not particularly limited as long as it is an arbitrary position including the clear area E ₂ . Therefore, in the second embodiment, the cutout width is limited. The details will be described below.

10A and 10B show fingerprint images G ₁ and G ₂ obtained at adjacent sample times t _i and t _{i + 1} when the finger is rotated right. Figure 10 (b), the center line C ₁ of the image G ₁
During right focal trough than a case where the center line C ₂ of the image G ₂ is present. Figure 10 (b), the image G ₂ centerline C ₂
Exists at a position on the right side, which is farther from the clear region of the image G ₁ . In the case as shown in FIG. 10A, the cutout for composition is center line C ₁ to center line C ₂ for image G _1.
And An example of combining the images thus cut out is shown in FIG.
It is shown in (c). Here, C ₃ is the center line of the fingerprint image of t _{i + 2 on} the right side of t _{i + 1} , and C ₄ is the center line of the fingerprint image of t _{i + 3} . In this way, if the center line interval W ₁ exists in the clear region, the fingerprint images obtained at each sample time are cut out at the adjacent center line intervals. On the other hand, FIG.
In (b), the cutout width from the image G ₁ is the centerline C ₁ ~
It is assumed that the width is up to C ₂ (W ₂ ), and the center line interval between the image G ₂ and the image G ₃ is about the width (W ₂ ). Therefore, FIG. 10D shows the result of cutting (cutting) with the center line interval width (W ₂ ) and combining (arranging). FIG.
The composite image in (c) is a composite image in which only clear images are cut out, and is an ideal rotated fingerprint image. On the other hand, the combined image of FIG. 10D is a defective image that cannot be handled as a rotating fingerprint image because it includes an unclear area and is a discontinuous (discontinuous) rotating fingerprint in the middle.

Therefore, the reference value W ₀ is set to the distance W between adjacent center lines.
If the centerline interval W is smaller than the reference value W ₀ (W ≦ W ₀ ), the image is cut out at the centerline interval W, and if W> W ₀ , the image is cut out. I decided not to. 10 (a) and 10 (b) are
Appears depending on the rotation speed of the finger. When the rotation speed is higher than expected, it is shown in FIG. 10B. When the rotation speed becomes low, the W shifts to W lower than W ₂ . FIG. 10A shows the time at the proper rotation speed. Reference value W
_{A value} of ₀ is a value such that a clear area can be included in the center line interval and clipping can be performed, and is set in advance.

FIG. 11 shows an arrangement example of the fingerprint image on the memory M, which is sampled along with the right rotation of the finger. One memory M _i is allocated for each sample point. Memory M _i
Is a two-dimensional memory composed of x and y coordinate systems, and x,
The coordinate system of y is the coordinate system x, y of the two-dimensional plane on the optical surface 2.
And corresponding. Therefore, the (x, y) coordinates on the optical surface correspond to the (x, y) coordinates of the memory M _i . Thus, the fingerprint image appearing on the optical surface is stored in the position on the memory M _i corresponding to the position on the optical surface.

Cutting is performed by using the memory M _i shown in FIG. First image G ₁ of the sampling time memory M ₁ obtained in t ₁
The x position of the center line C ₁ is calculated for. For sample time t ₂ adjacent after obtaining the image G ₂ in the memory M _2, and calculates the x position of the center line C ₂ of the image G _2. Then, a difference (interval) W between the x position of C _{1 and} the x position of C ₂ is obtained, and it is determined whether this is larger or smaller than the reference value W ₀ . W ≦ W ₀
If so, the image G _{1 is} cut out at the interval W, starting from the x position of the center line C ₁ . Thereafter, the sample time t is updated and the same processing is performed. If the cutout images are arranged continuously, it is a composite image, and a composite image (entire image of the rotating fingerprint) is obtained. An example of composition by this continuous arrangement is shown in FIG.

Incidentally, FIG. 11 shows a memory M having the number of samples.
_i is required. However, as can be seen from FIG. 11, the memory M _i for each sample time has many blank areas and uses too much memory. Therefore, in practice, the memory M _i that is considered to have the maximum capacity for one fingerprint image is set, and the center lines C ₁ and C ₂ and the interval W are obtained by software processing,
Cutting out at an interval W is performed under the condition of W ≦ W ₀ . Also, 1
It is possible to use only one memory.

A third embodiment of the present invention will be described. In the second embodiment, the cut was centerline spacing adjacent, was so cut out by the permissible maximum sample interval value W _m is a the third embodiment. With the maximum permissible sample interval value W _m , cutting is not discontinuous,
It is also the maximum allowable value for cutting out a clear image. Then, based on the center line, an image having a width of W _m is cut out from the center line to the right. On the other hand, when clipping is performed with such a constant interval value W _m , the clipped image may become discontinuous if the rotation speed of the finger is faster than expected. Therefore, the center line C ₁ for the current captured image G ₁
The position of the center line C ₂ after determined in advance by performing a cut, the adjacent center line C ₂ of the next image G ₂ with a width of W _m from,
The previous image G ₁ (center line C ₁ + W _m ) is compared to determine whether (center line C ₁ + W _m ) exceeds center line C ₂ . And if it exceeds C ₂ , it is considered that it is not a discontinuous cutout, and the previous cutout is allowed. If C ₂ is not exceeded, an alarm will be output indicating that there is a discontinuous portion between the previous cutout and the current cutout, the fingerprint rotation will be centered, the finger rotation from the beginning will be performed again, and the capture will be repeated. Measure

Next, a fourth embodiment of the present invention will be described. This embodiment aims at correcting rotational distortion. Examples of the strains described in the above problems include strains in a direction orthogonal to the rotation direction and rotation strains around an axis in a direction orthogonal to the rotation direction, but these have large curvatures of the fingertips. There are mainly causes. This is shown in FIG. A region E in FIG. 12 shows a portion having a large radius of curvature. When a finger is rotated including the fingertip having such a large curvature, the rotation is performed linearly from left to right, but the fingertip does not move along the straight line but moves along the curvature. This causes rotational distortion. Of course, it may occur slightly due to other ways of moving the finger, differences in the shape of the finger, and the like. Therefore, before and after the center line C that is the composite boundary line, it is decided to compare the images before and after the center line C to detect whether or not a deviation has occurred. Further, if the occurrence of the deviation is detected, the deviation due to the large curvature is the main cause. Therefore, the deviation is corrected by linearly or non-linearly converting the shape of the two-dimensional image by the affine transformation or other method.

Next, an embodiment of such deviation detection and deviation correction is shown in FIGS. FIG. 13 is a conceptual diagram of deviation detection. The image sampled at time t ₁ is G ₁ (FIG. 10).
(A)), the image sampled at time t ₂ adjacent to t ₁ is G
₂ (Fig. 10 (b)). The x coordinate position of the center line C ₁ of the image G ₁ is x ₁ , and the x coordinate position of the center line C ₂ of the image C ₂ is x ₂ . In the image G ₁ , an oblique line image in the section of x _{1 to} x ₂ ,
In the image G ₂ , the hatched image in the section of x _{2 to} x ₃ becomes the cutout image, and by arranging this continuously,
The composite image of (C) is obtained. Therefore, in order to detect the shift, the x coordinate x ₂ of the image G ₁ is in the vicinity of ± ΔW (x
₂ −ΔW to x ₂ + ΔW) and x for the image G _2.
The amount of displacement between the coordinate x ₂ and the vicinity of ± ΔW (section between x ₂ −ΔW and x ₂ + ΔW) is calculated. If there is no deviation, the two match. If there is a deviation, a discrepancy corresponding to the deviation amount appears. FIG. 13D shows this in relation to the y-axis-deviation amount. It is shown that a relatively large amount occurs at the fingertip with a large radius of curvature shown in FIG.

FIG. 14 shows a deviation correction method. Therefore, from FIG. 13 (d), the deviation start point P ₂ , the arbitrary point P ₃ of the fingerprint image G ₁ , the point P ₄ of the fingerprint image G ₂ corresponding to P ₃ , and the fingerprint image at the position x ₂ Find the upper end point P ₁ of G ₁ . Then, a tangent line tangent to the upper side of the fingerprint image G ₁ is obtained from the point P _1, and an intersection point P ₅ between this and the straight line C _{2 at} the position x ₂ of the image G ₂ is obtained. To correct the deviation, a two-dimensional shape conversion is performed on the triangle surrounded by the points P ₁ to P ₂ and P ₅ so that the points P ₁ and P ₂ are fixed and the points P ₃ and P ₄ are the same point. To achieve in.

FIG. 1 shows an embodiment of a rotary fingerprint detecting apparatus of the present invention. In this embodiment, the light source 4, the optical prism 1,
In addition to the optical system including the two-dimensional sensor 5, the two-dimensional sensor 5
A combination processing unit 1 that combines the captured fingerprint images detected by
5 and a monitor 14 as display means. The synthesis processing unit 15 includes an AD converter 7, a delay circuit 8, a center detection circuit 9, a synthesis memory 10, a write control circuit 11, a DA converter 12, and a distortion correction circuit 13. The operation will be described.
The fingerprint image input from the two-dimensional sensor 5 is the A / D converter 7
Are converted into multi-tone digital data. The converted data is sent to the center detection circuit 9. The center detection circuit 9 obtains the position x of the center line C of each sample fingerprint image.
The center position x may be determined by centering the center of gravity of the fingerprint image in the x direction, or by binarizing the original fingerprint image and determining the center of gravity. Further, the center in a band-shaped range having the fingerprint area may be obtained. Any method may be used as a method of obtaining the center.

The center detection circuit 9 sends the coordinates of the obtained center line C to the write control circuit 11. The center detection circuit 9 and the write control circuit 11 require processing time for a certain time (generally, one frame time of an image), so that the digital data from the A / D converter 7 is time-adjusted by the delay circuit 8 and synthesized. Sent to the memory 10 for use. The synthesizing memory 10 performs synthesizing according to a command from the write control circuit 11. After the combination, it is displayed as a rotary fingerprint on the monitor 14 via the DA converter 12. If distortion correction is necessary, distortion correction circuit 1
In 3, the distortion detection is corrected. If necessary, this correction result is displayed on the monitor. The delay circuit 8 may be removed and a buffer memory may be provided instead. In this case, the function of the delay circuit 8 is achieved by adjusting the timing of reading / writing from the buffer memory. This buffer memory is, for example, a memory M _{i as shown} in FIG. Further, the memory M _{i in} FIG. 11 may be regarded as the composition memory 10.

FIG. 2 shows a synthesis processing flow executed mainly by the write control circuit 11. When the rotation fingerprint input is started, the internal register is initialized in Flow 100,
It also initializes the synthesis memory. The next flow 101 is a part for determining whether or not the capturing is completed. If the capturing is clockwise, it is detected that the fingerprint image on the right side is input and the center line transmitted is hardly advanced for a certain period, and this is detected. If so, it is considered that the fingerprint collection work has been completed by stopping the rotation, and the acquisition is completed.

[0032] In flow 102, calculates the center position x _i of the captured image G _i, latches it. Flow 103
Then, it is checked whether the captured image G _i is the start image of the rotating fingerprint. If it is the start image, the flow proceeds to flow 104, and all the images on the left side of the center position x _i obtained in flow 102 are written in the synthesis memory 10. The writing position is the left image corresponding position (0 to x _i ).

[0033] If the starting image, move to the flow 105, the last captured image G _i-1, from the center position x _i-1 in the latched, the image clipping over an interval to the center position x _i of G _i Do that and put it in synthetic memory x
Write at positions _{i-1 to} x _i . In the frame 106, the center position interval W = x _i −x _i−1 is obtained, and it is checked whether it is within the reference value W ₀ . If W> W ₀ , an alarm is generated in flow 107, fingerprint collection is stopped, and collection is performed from the beginning. This flow 107 is flow 1
It may be provided between 03 and 105. If W ≦ W ₀ , step 10 is executed to fetch the next image G _{i + 1.}
Fly to 1. Also, when the image is the start image of the flow 104, the process jumps to step 101.

FIG. 15 shows a synthesis processing flow for realizing the third embodiment. In this synthesizing process, the maximum permissible interval Wm is set, and the image in the section from the center position x _i to (x _i + W _m ) is automatically cut out from the captured image G _i . Then, (x _i + W _m ) is compared with the center position x _{i + 1 of the} next captured image G _{i + 1} to obtain (x _i + W _m ).
If ≧ x _{i + 1} , the cutout from G _i from x _i to (x _i + W _m ) is regarded as valid, and the subsequent combining process is continued. If (x _i + W _m ) <x _{i + 1} , the cut-out image is discontinuous, so it is regarded as improper cut-out, the synthesis process is interrupted, an alarm is sounded, and re-sampling is performed.

The processing of FIG. 15 will be described based on the above concept. When the rotation fingerprint input is started, the internal register is initialized in flow 100, and the synthetic memory is also initialized. The next flow 101 is a part for determining whether or not the capture is finished. If the capture is clockwise, the fingerprint image on the right side is input and it is detected that the center line sent is hardly advanced for a certain period and the capture is completed. to decide. Next, in the flow 108, the cutting range of the previous fingerprint image is saved, and the center line sent this time is saved as the cutting start place. The cut end point will be saved by adding the maximum sample interval W _m to the starting point cut. In flow 109, it is determined whether or not the fingerprint image is the first one after the input of the rotary fingerprint. In the case of the first fingerprint image, since the portion opposite to the rotation direction of the center line is also written in the flow 110, the width W _m from the coordinate 0 on the rotation start side
Are written in the composition memory 10. In the case of a fingerprint image in the middle, in flow 111, from the center position x _i of the captured image G _{i to} the coordinates (x _i + W _m ) are arranged in the composition memory and combined. Next, in step 112, the previous write end coordinate and the current write start coordinate are compared, and if (x _i-1 + W _m ) is smaller than x _i, it means that the sample interval is wider than the allowable sample interval. Output an alarm. In addition to this flow, the writing control circuit 10 outputs the coordinates of the center line of each fingerprint image and the coordinates before and after the center line for distortion detection to the distortion correction circuit 12.

The writing in the flow 111 overlaps between adjacent images G _i and G _{i + 1} at a part of the x coordinate as shown in FIG. That is, interval time of cutting out the section preceding the clipped image G ₁₁ is, x ₁ ~ of 0 to w _m in FIG. 16 (x ₁ + W _m) is G _21. Since the section of x _{1 to} W _m is overwritten on the already-written image G ₁₁ when writing the image G ₂₁ , there is no problem due to this duplication. It should be noted that there is no x-coordinate overlap in the combining process of FIG.

When the composition is completed in the composition memory 10 in FIG. 1, the result is sent to the distortion correction circuit 13 for distortion detection and correction. Also, during the synthesis and the synthesis result is D / A
It is sent to the monitor 14 through the converter 12 and can be monitored at any time. The distortion correction circuit 13 is generally composed of a CPU and a memory, and the combined output of the combining memory 10 and
Before and after the center line of the delay circuit 8 specified by the center line of each fingerprint image from the writing control circuit 11, the coordinates before and after the center line for distortion detection, and the coordinates before and after the center line for distortion detection. The fingerprint data is input to correct distortion and erroneous synthesis. In this way, it is possible to input a clear and erroneously synthesized rotating fingerprint.

[0038]

According to the present invention, it is possible to detect a rotating fingerprint that is clear and does not cause erroneous synthesis.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of a rotary fingerprint detecting device of the present invention.

FIG. 2 is a diagram showing an example of a flowchart of a combining process of the present invention.

FIG. 3 is a diagram showing a conventional rotary fingerprint input system.

FIG. 4 is a diagram showing reflection of a conventional rotary fingerprint.

FIG. 5 is an explanatory diagram of conventional synthesis of a rotating fingerprint.

FIG. 6 is a composite view of a conventional rotary fingerprint.

FIG. 7 is an explanatory diagram of a problem associated with detection of a rotating fingerprint.

FIG. 8 is an explanatory diagram of rotational distortion.

FIG. 9 is an explanatory diagram of rotational distortion.

FIG. 10 is an explanatory diagram of a combining process of the present invention.

FIG. 11 is an explanatory diagram of a synthesizing process of the present invention on a memory.

FIG. 12 is an explanatory diagram of rotational distortion of a finger.

FIG. 13 is an explanatory diagram of rotational distortion detection according to the present invention.

FIG. 14 is an explanatory diagram of the rotational distortion correction of the present invention.

FIG. 15 is a diagram showing an example of a flowchart of a process for synthesizing a rotated fingerprint according to the present invention.

16 is a diagram showing a partial overlap on the memory in the processing flow of FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 prism 2 optical surface 3 finger 4 light source 5 two-dimensional sensor 7 AD converter 8 delay circuit 9 center detection circuit 10 synthesis memory 11 writing control circuit 12 DA converter 13 distortion correction circuit 14 monitor 15 synthesis processing unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Takayuki Komamura 3-2-1, Sachimachi, Hitachi City, Ibaraki Prefecture Inside Hitachi Engineering Co., Ltd.

Claims (10)

[Claims] 1. A rotating fingerprint detecting method in which a finger is rotated on an optical surface, and this is measured to detect an entire image of a rotating fingerprint, a rotating fingerprint image on the optical surface is measured as an image, and each image thus obtained The method for detecting a fingerprint of rotation, wherein an image having a predetermined width in the direction of rotation is cut out into strips from a centerline orthogonal to the direction of rotation of the image, and the cutout images are combined to obtain an entire image of the rotated fingerprint. 2. A rotating fingerprint detecting method, which comprises rotating a finger on an optical surface and measuring the image to detect an entire image of a rotating fingerprint, wherein the rotating fingerprint image on the optical surface is measured as an image, and each image thus obtained. About, the image from the centerline position orthogonal to the rotation direction of the image to the centerline position of the next image along the rotation direction is cut into strips, and the cutout images are combined to create the entire image of the rotated fingerprint. A method for detecting a rotating fingerprint that is obtained. 3. A rotating fingerprint detecting method for rotating a finger on an optical surface to measure the whole image of a rotating fingerprint by measuring the finger, wherein the rotating fingerprint image on the optical surface is measured as an image, and each image thus obtained. A method for detecting a fingerprint of a rotation, in which an image with a maximum sample width (W _m ) is cut out in a strip shape from a center line orthogonal to the rotation direction of the image, and the cut out images are combined to obtain an entire image of the rotation fingerprint. . 4. The rotary fingerprint detection according to any one of claims 1 to 3, wherein the presence or absence of rotational distortion of the finger is detected at the composition boundary, and if the rotational distortion occurs, the distortion is corrected. Method. 5. The rotating fingerprint detecting method according to claim 2, wherein
A rotary fingerprint detecting method in which when the cutout section is larger than a predetermined first reference value (W ₀ ), the cutout is notified as a defect. 6. The rotating fingerprint detecting method according to claim 3,
A rotating fingerprint detecting method in which the cutout is notified as a defect when the maximum sample width W _m does not reach the center line position of the next adjacent image. 7. A rotating fingerprint detecting device for rotating a finger on an optical surface and measuring the whole image by detecting the rotating fingerprint, and a means for capturing the rotating fingerprint image on the optical surface and each screen thus obtained. And a means for cutting out an image of a predetermined width in a strip shape, which is oriented in the rotation direction from a center line orthogonal to the rotation direction of the image, and means for synthesizing the cut-out images and detecting the entire image of the rotated fingerprint. Rotating fingerprint detector. 8. A rotary fingerprint detecting device for rotating a finger on an optical surface and measuring the fingerprint to detect a whole image of a rotary fingerprint, a means for capturing a rotary fingerprint image of an optical surface, and each screen thus obtained. , A section from the centerline orthogonal to the rotation direction of the image to the centerline position of the next image along the rotation direction, and means for cutting the image into strips, and combining the cutout images, the entire image of the rotated fingerprint A rotary fingerprint detecting device, which comprises: 9. A rotating fingerprint detecting device for rotating a finger on an optical surface and measuring the whole by rotating the finger to detect the entire image of the rotating fingerprint. , A means for cutting out an image having a maximum sample width (W _m ) in the shape of a strip from the center line orthogonal to the rotation direction of the image, and combining the cut out images to detect the entire image of the rotated fingerprint. A rotary fingerprint detecting device comprising: 10. The detection device according to claim 7, further comprising means for detecting the presence / absence of rotational distortion of the finger at the synthesis boundary and correcting the rotational distortion if it occurs. Rotating fingerprint detection device.