CN100362526C - Image input apparatus - Google Patents
Image input apparatus Download PDFInfo
- Publication number
- CN100362526C CN100362526C CNB2005101170231A CN200510117023A CN100362526C CN 100362526 C CN100362526 C CN 100362526C CN B2005101170231 A CNB2005101170231 A CN B2005101170231A CN 200510117023 A CN200510117023 A CN 200510117023A CN 100362526 C CN100362526 C CN 100362526C
- Authority
- CN
- China
- Prior art keywords
- data
- sensor
- image
- finger
- mentioned
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Image Input (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Collating Specific Patterns (AREA)
Abstract
An image input apparatus for sequentially obtaining a plurality of partial images which are portions of an image of an object, by relatively moving the object and a sensor having a plurality of pixel circuit units, including a drive circuit which effects control so as to store photoelectric charge en bloc at photoelectric conversion units respectively included in the plurality of pixel circuit units, then blanket-transfer optical signals of the photoelectric conversion units respectively included in the plurality of pixels to memory units included in the corresponding pixel circuit units, and then read the photoelectric signals, and noise signals generated in the pixel circuit units in this order from the pixel circuit units.
Description
The present invention is that application number is 03104571.5, the applying date is on February 18th, 2003, denomination of invention is the division of the patented claim of " image-input device ".
Technical field
The present invention relates to discern the image-input device and the fingerprint identification device of characters such as original copy, bar code, fingerprint and image etc.
Background technology
In the prior art, there is such formation: in the image-input device of character such as identification original copy, bar code and image etc., send by hand hand scanner etc. is moved, carry out image and read on original copy etc.This formation can implement device miniaturization, portability, still, the hand that has a hand scanner send velocity variations and causes the situation of anamorphose.
In addition, there is such image-input device: finger is directly touched on the semiconductor transducer (capacitance type and optically read type etc.) of two dimension, read fingerprint, carry out finger print identifying.In recent years, such fingerprint sensor is being gazed at aspect enterprise or individual's security personnel's the reinforcement, and, utilizing mobile phone and carrying in the ecommerce etc. of equipment, if it is small-sized, in light weight, low-cost that fingerprint sensor becomes, will expect the market expansion that leaps.But present fingerprint sensor is to constitute with the big or small identical size with finger, and therefore, it is large-scale that size sensor has to become.
Semi-conductive price is decision uniquely by chip yield, and therefore, size causes high price greatly, in the time of in being loaded into mobile phone, has problems on cost.And large-scale fingerprint sensor is not suitable for small-sized portable equipment.Therefore, if make sensor become small-sized, then on cost, be favourable in shape.As solution, if adopt optically read type, constitute optical sensor with line-array sensor, can make the sensor miniaturization.But, in the case,, just must make line-array sensor and finger relative motion in order to read fingerprint, therefore, with read by the image that hand scanner carried out identical, the distortion of the fingerprint image that reads, authentication precision significantly reduces.
Therefore, as the existing known example of improving this shortcoming, there is Japanese patent gazette to specially permit No. 3198810.It is: the hand the when character of judgement original copy reads send speed, and this judged result is fed back to the sequential generating means, comes the correcting image distortion.
In No. 3198810 disclosed technology of above-mentioned special permission, under the situation of exactissima diligentia in original copy reads and mobile lentamente original copy or hand, perhaps can carry out the authentication of image with low resolution.But can not be used for needing high resolving power, require the such purposes of fingerprint sensor of high authentication precision.
Summary of the invention
The purpose of this invention is to provide the image-input device that a kind of face sensor with fewer pixel column reaches the high repeatability of reading images.
In order to achieve the above object, in one embodiment of the invention, a kind of image-input device is provided, by subject and the sensor that comprises a plurality of pixels are relatively moved, divide the image of above-mentioned subject and obtain a plurality of parts of images respectively successively as the part of the image of above-mentioned subject, so that have overlapping between resulting above-mentioned a plurality of parts of images, it is characterized in that, comprise: control circuit, according to the comparison of a plurality of above-mentioned parts of images data, control obtains the timing of a plurality of above-mentioned parts of images data that will synthesize; And image synthesizer, the lap of above-mentioned a plurality of parts of images that correction is exported from the sensor and to carry out image synthetic, wherein, above-mentioned control circuit comprises commutation circuit and testing circuit, described commutation circuit, switching first is read mode and second and is read mode, described first reads mode utilizes a plurality of pixels of the sensor to carry out image to read, described second reads mode utilizes carrying out image and read than reading a plurality of pixels that the pixel count that uses in the mode lacks above-mentioned first of the sensor, described testing circuit, for the subject that relatively moves with the sensor, relatively reading mode with above-mentioned first reads resulting parts of images data and after this reads resulting parts of images data with the above-mentioned second reading mode of getting, reading resulting parts of images data with the above-mentioned second reading mode of getting and reading mode with above-mentioned first and read when between the resulting parts of images data corresponding to data being arranged when above-mentioned testing circuit detects, by above-mentioned commutation circuit, read mode from above-mentioned second and switch to above-mentioned first and read mode.
Other purposes of the present invention, feature can be come to understand by following instructions and accompanying drawing.
Description of drawings
Fig. 1 is the system diagram of an embodiment of image-input device of the present invention;
Fig. 2 A is the figure that moves that schematically represents sensor and finger, and Fig. 2 B is the figure of concise and to the point formation of the fingerprint reading device of expression optical profile type;
Fig. 3 be the word of identification fairly simple " M " shape of expression shape be subject as the time the synoptic diagram of view data;
Fig. 4 is the figure of the circuit block of expression sensor;
Fig. 5 is the figure of expression sensor pixel portion;
Fig. 6 is the sequential chart of the action sequence of expression sensor;
Fig. 7 is the system diagram of an embodiment of image-input device of the present invention;
Fig. 8 A is the figure that moves that schematically represents sensor and finger, and Fig. 8 B is the figure of concise and to the point formation of the fingerprint reading device of expression optical profile type;
Fig. 9 is used for illustrating that the one part of pixel alignment of sensor reads the figure of action;
Figure 10 is used for the process flow diagram of key diagram as input action;
Figure 11 be the word of identification fairly simple " M " shape of expression shape be subject as the time the synoptic diagram of view data;
Figure 12 is the figure of the circuit block of expression sensor;
Figure 13 is the figure of expression sensor pixel portion;
Figure 14 is the sequential chart of the action sequence of expression sensor;
Figure 15 is the sequential chart of the action sequence of expression sensor;
Figure 16 is the system diagram of an embodiment of image-input device of the present invention;
Figure 17 is the system diagram of an embodiment of image-input device of the present invention;
Figure 18 A is the figure that moves that schematically represents sensor and finger, and Figure 18 B is the figure of concise and to the point formation of the fingerprint reading device of expression optical profile type;
Figure 19 A, 19B are the figure that reads action that is used for illustrating sensor;
Figure 20 be the word of identification fairly simple " M " shape of expression shape be subject as the time the synoptic diagram of view data;
Figure 21 is the figure of the circuit block of expression sensor;
Figure 22 is the figure of expression sensor pixel portion;
Figure 23 is the sequential chart of the action sequence of expression sensor;
Figure 24 is the sequential chart of the action sequence of expression sensor;
Figure 25 is the figure that reads action that is used for illustrating sensor;
Figure 26 is the sequential chart of the action sequence of expression sensor;
Figure 27 be the word of identification fairly simple " M " shape of expression shape be subject as the time the synoptic diagram of view data;
Figure 28 be the word of identification fairly simple " M " shape of expression shape be subject as the time the synoptic diagram of view data.
Embodiment
Use accompanying drawing to come embodiments of the invention are described below.In the following description, mainly be read as example with the fingerprint of the high-precision image regeneration of needs and describe, still, the present invention is not limited in this example, also can be used for original copy and read with bar code etc.In addition, move finger with respect to sensor here, still, also can sensor be moved with respect to subject as subject.
Below embodiment 1 is described.
Fig. 2 A is the figure that moves that schematically represents sensor and finger, and Fig. 2 B is the figure of concise and to the point formation of the fingerprint reading device of expression optical profile type.In the present embodiment, the sensor 10 of usage level pixel 540 pixels, vertical pixel 20 pixels.When carrying out finger print identifying, for example, using Vertical number of pixels usually is about 500 pixels, and therefore, in contrast to this, the size sensor of Shi Yonging is very little in the present embodiment.In addition, in the present embodiment, not to utilize sensor to read all images, but read its a part of image that therefore, Vertical number of pixels can be the pixel count that fully reads a part of image.On sensor 10 (as required; guard blocks such as configuration sheet glass or fibre plate on sensor 10) place finger; in the present embodiment; shown in Fig. 2 A like that; make finger on for the vertical direction of the length direction of sensor 10, move (representing), reading images successively with the arrow among Fig. 2 A.
Shown in Fig. 2 B like that, from the sidepiece of finger (can be go up direction, oblique upper to, down direction, tiltedly descend direction, in order to seek the miniaturization of device, preferably from being from sidepiece) by LED 1 irradiates light as light irradiation device, in finger, light scattering is seen through, this light incides the sensor 10 from finger, obtains fingerprint image.
Fig. 1 is the system diagram of an embodiment of image-input device of the present invention.Sensor 10 is exported the photographic images of the part of finger successively.There is the part that repeats in the image of the part of the finger that is read respectively by sensor 10.Along with the translational speed of finger is slack-off, it is many that the ratio of the part of repetition becomes.Sensor 10 carries out drive controlling by not shown driving circuit.
Picture signal from sensor 10 temporarily is kept in the storer 12, perhaps, is directly inputted in the unique point extraction unit 11, and the unique point of the image of a part of pointing is extracted out.Extract out for unique point, be described in addition.Unique point is stored in the storer 13 by every pixel column arbitrarily each frame.For example, first frame is written into memory area 1, the second frame and is written into memory area 2 ..., the n frame is written into memory area n.The unique point of each frame compares in unique point comparing section 14, is calculated translational speed or the velocity variations of pointing on the position of the corresponding to pixel column of unique point by speed detection unit 16.Simultaneously, determine that by overlapping detection unit 15 the overlapping of image of interframe is the overlapping region of image, revise (correction) in image correction portion (image synthesizes portion) 17, making does not have the overlapping of image, shows on display part 18.This image is stored in the data preservation portion 20 when login, when authentication, compare with the data of in authentication department 19, being preserved, with the image of the finger that is corrected consistent, the inconsistent authentication result of indicator gauge first finger line together, under inconsistent situation, that points once more moves on display part 18, under corresponding to situation, enter the purpose (for example, allow the use of mobile phone, can converse) of this equipment.
For the unique point extracting method, be generally following three kinds of methods: extract out from fingerprint image carina fingerprint characteristics such as structure (end points and take-off point) method (for example, the Marchal relations act), image (figure) matching method of the similarity of evaluation map picture, laterally decomposing fingerprint image meticulously on (with the vertical direction of length direction of finger), the frequency analysis method that the deep or light difference of the carina of each image is carried out frequency analysis.
Fig. 3 be the word of identification fairly simple " M " shape of expression shape be subject as the time the synoptic diagram of view data.Data A represents the view data of the subject picture of t1 sometime, and data B represents that (view data of＞subject picture on t1), data C represents the view data of lap for t2 constantly.By the view data of comparing data A and the view data of data B, the data C of the lap of recognition image is modified to the view data of the subject picture of " M " shape.
Shown in the synoptic diagram of Fig. 3 like that, under the situation of the fairly simple subject picture of the identification of shape, the method for the feature of data A and data B as a comparison is preferably the images match method of the similarity of evaluation map picture.Resemble under the situation of the such more complicated of fingerprint, if relatively more a part of fingerprint, then the processing of frequency analysis method is fairly simple.The Marchal relations act detects the end points and the take-off point of carina, and therefore, preferably sensor area is bigger.
Fig. 4 and Fig. 5 are the figure of a configuration example of expression sensor 10, and Fig. 4 is the figure of the circuit block of expression sensor, and Fig. 5 is the figure of expression sensor pixel portion.
In Fig. 4, the 30th, the photoelectric conversion department of reading images, the 31st, the vertical scanning circuit of the scan line of selection photoelectric conversion department 30, the 32nd, the temporary transient temporary transient preservation storer of preserving from the signal of photoelectric conversion department 30, the 33rd, the signal transmission from temporary transient preservation storer 32 is controlled to the differential amplifier of back level and the horizontal scanning circuit of gain variable amplifier (PGA) 34.Signal from PGA34 is transformed to digital signal by analog digital converter (A/D transducer) 36, imports the unique point extraction unit 11 shown in Figure 1 of back level.Vertical scanning circuit 31, the control signal of temporarily preserving storer 32, horizontal scanning circuit 33, PGA 34, A/D transducer 36 origin self-timing generators 35 are controlled.
Then, use Fig. 5 that photoelectric conversion department shown in Figure 4 30 and temporary transient formation of preserving storer 32 are described.Photoelectric conversion department is made of the pixel amplifier form that is called as cmos sensor as used herein.Each pixel is formed by 6 elements.That is, pixel by photodiode PD, remove the reset switch M of the residual charge of photodiode PD
V, be used for the signal charge of photodiode PD is sent to the transmission switch M of the amplifier of back level
TX, be used to remove the reset switch M of residual charge of the gate portion of amplifier (buffer transistor)
RES, as the buffer transistor M of the amplifier of the signal voltage of buffered gate portion
SF, buffer transistor M
SFSignal output control to switch M on the vertical signal line
SELConstitute.M
1And M
2Be the capacitor C that is used for send temporary transient preservation usefulness to from the signal of vertical signal line
S, C
NSwitch.M
3And M
4Be to be used for handle from capacitor C
S, C
NSignal S
1With noise N
1Send the switch of next differential amplifier (not shown) respectively to through the horizontal output line.Be used to sum up the switch M that resets
V, M
RESAnd the transmission switch M that is used to sum up transmission
TXBe equivalent to the blanket exposure device of whole pixels.Be to be reset to by transmitting switch M between the blanket exposure period of pixel all from whole pixels
TXDuring till the signal that produced transmits and finishes, by lighting LED 1, and can access the fingerprint image of finger during this period.
The signal of photodiode PD is transmitted to capacitor C
S, pixel amplifier (buffer transistor M
SF) reset noise be transmitted to capacitor C
N
Use Fig. 6 to come the action sequence of sensor is described below.
The basic driver sequential of Fig. 6 is: after whole pixels of sensor were summed up removing comprehensively, the LED 1 shown in Fig. 2 (B) was lighted, and the part of fingerprint is read.This signal is by blanket amplifier gate portion (the buffer transistor M that sends each pixel to
SFGate portion).The signal of gate portion is transmitted to capacitor C in every row
SPreserve.Then, gate portion is reset, and its noise is stored in capacitor C
NIn.Signal and noise are sent to the differential amplifier of back level simultaneously, remove denoising from signal.This action is carried out in every row repeatedly.
Use Fig. 6 to describe above-mentioned action in detail below.At first, when the finger on the sensor 10 was detected, during T1, the action of sensor began, from timing generator 35 output control signal φ
StartAccept control signal φ
Start Vertical scanning circuit 31 during T2 at first signal psi
C1, φ
C2 (n), φ
C2 (n+1)... as the H level, connect switch M
V, M
RES, gate portion (the buffer transistor M of removing photodiode PD and amplifier in whole pixels
SFGate portion).Then, make signal psi
PIn whole pixels is the H level simultaneously, connects and transmits switch M
TX, at this moment, LED 1 is lighted (φ
LEDBe the H level), shine on the mobile finger, carry out light-to-current inversion from the reflected light of finger by the photodiode PD of each pixel, to gate portion (the buffer transistor M of amplifier
SFGate portion) send signal from photodiode PD.Fingerprint image carries out reading of 20 row.And, make transmission switch M
TXUnder the state that turn-offs, during in the T3, light LED 1, then, connect and transmit switch M
TX, and can transmit signal.Control so that during the lighting of LED 1 between (during the rayed) and the exposure period that is produced by the blanket exposure device of whole pixels consistent or be between this exposure period in, thus, can reduce power consumption.The output signal of amplifier is by making signal psi
Sel (n)And φ
SBecome the H level, and connect switch M
SELAnd M
1, thus, be stored in capacitor C
SIn.
Then, make signal psi
C2 (n)Become the H level, make switch M
RESConnect, make amplifier gate portion (buffer transistor M
SFGate portion) reset.Then, make signal psi
Sel (n)And φ
NBecome the H level, and connect switch M
SELAnd M
2, thus, reset noise is transmitted to capacitor C
NAt this moment, photodiode PD for prevent because of ambient light etc. by the leakage of electric charge that charge excess produced to gate portion, make signal psi
C1Become the H level, switch M
VConnect, be controlled on the set potential.Be stored in capacitor C
S, C
NInterior signal and noise are sent to the back level successively and handle.Then, when the signal of the capable pixel of n and noise are read successively, the output signal of the amplifier of the capable pixel of the n+1 that follows is sent to capacitor C in every row
S, reset noise is sent to capacitor C
N, carry out the action of reading that transmits processing in the back level.Like this, carry out reading of each pixel column.T4 whole picture element signals read into during the level of back during this time.When on detecting sensor 10, not pointing, from timing generator 35 output signal φ
End, the action of stop sensor.
At this, the establishing method of reading the time of whole pixels is important.That is, need the top speed that moves of anticipation or default finger, the frame speed of decision sensor.When frame speed was slow, because finger is mobile, fingerprint was fuzzy, perhaps, when frame speed is too fast, under the mobile slow situation of finger, must prepare a plurality of video memories.
In the present embodiment, it is square that Pixel Dimensions is set at 30 μ m, and the highest translational speed of finger is set at 1cm/0.1sec.At this moment, if during LED lights be 28 μ sec, the mobile of finger during this period is 2.8 μ m, and the fuzzy part of image or the sampling error of interframe are below 1/10 of max picture size.When image overlapping was minimum 5 pixels, about 1.4msec carried out repeatedly in 150 μ m, and frame rate is about 714 frame/seconds.Control signal φ
ReadBe about 1.4msec during this time, therefore, signal output frequency is about 8MHz.In the time of during further shortening LED lights, can reduce the sampling error of image, improve the fingerprint repeatability after image synthesizes, therefore, can improve authentication precision.
If as prerequisite, it is synthetic that driving sensor carries out image described later, and the unique point extraction method just can be realized satisfiable authentication with frequency analysis method, images match method, relations act.But (authentication precision that needs) as required sets unique point extracting method or frame speed.
In addition, method as fingerprint detection, except to the finger irradiates light, with its reflected light of light sensors or see through light (perhaps scattering sees through light) and detect outside the mode of optical profile type of fingerprint, also exist finger is placed on the semiconductor transducers such as piezoelectric element, electrostatic condenser detecting element, detector unit, detect the semiconductor transducer mode of fingerprint by detected pressures, electric capacity, temperature, can use any detection method in the present invention.But, in the present invention, because photographic images makes the part of live image repeat, so require the translational speed fast sensor of the reading speed of sensor 10 than finger.
As described above, according to present embodiment,,, can realize miniaturization so cost is low owing to use less sensor.And, can reduce power consumption, carry out high-precision authentication, and can in portable equipment, use.
Below embodiment 2 is described.
Fig. 8 A is the figure that moves that schematically represents sensor and finger, and Fig. 8 B is the figure of concise and to the point formation of the fingerprint reading device of expression optical profile type.In the present embodiment, the usage level pixel count is 540 pixels, and Vertical number of pixels is the sensor 10 of 20 pixels.When carrying out finger print identifying, for example, using Vertical number of pixels usually is about 500 pixels, and therefore, in contrast to this, the size sensor of Shi Yonging is very little in the present embodiment.And, in the present embodiment, not to read all images, but read its a part of image that therefore, Vertical number of pixels can be the pixel count that fully reads a part of image by sensor.On sensor 10 (as required; guard blocks such as configuration sheet glass and fibre plate on sensor 10) place finger; in the present embodiment; as shown in Figure 2; make finger on for the vertical direction of the length direction of sensor 10, move (representing), reading images successively with the arrow among Fig. 2.
Shown in Fig. 8 B like that, from the sidepiece of finger (can be go up direction, oblique upper to, down direction, tiltedly descend direction, in order to seek the miniaturization of device, preferably from being from sidepiece) by LED 1 irradiates light as light irradiation device, in finger, light scattering is seen through, this light incides the sensor 10 from finger, obtains fingerprint image.
In the present embodiment, using Fig. 8 A, 8B explanation based on before the reading of the fingerprint image of the shift action of finger, as shown in Figure 9, do not use sensor 10 whole pixel alignments (this is 20 the row alignments) all, and use a plurality of pixel alignments of a part to come read output signal, judge on the sensor 10 and have or not finger, after can confirming that finger is configured on the sensor 10, whole pixel alignments of use sensor 10 enter reading in based on the fingerprint image of the shift action of finger.In this constitutes, when judging having or not of pointing, do not scan whole pixel alignments, therefore, can seek the low consumption electrification of system.
Fig. 7 is the system diagram of an embodiment of image-input device of the present invention.At first, as shown in Figure 9, use a plurality of pixel alignments of the part of sensor 110 to come read output signal, judge on the sensor 110 to have or not finger.Promptly, drive a plurality of pixel alignments of the part of sensor shown in Figure 9 110 repeatedly, and obtain data A1, data A2, data A3 ... (each data representation is from the output data of a plurality of pixel alignments of a part), these data through storeies 112 or be directly inputted to unique point extraction unit 111.In unique point extraction unit 111, when detecting unique point that the state of not extracting (finger place) out from the unique point of finger becomes finger when being drawn out of the state of (having placed finger), unique point extraction unit 111 sends switching signals to switching part 121, and switching part 121 uses the signal of reading beginning based on the fingerprint image of the shift action of finger of whole pixel alignments shown in Figure 2 to sensor 110 output indications.
After there be moving of being identified and pointing in the finger on the sensor 110, sensor 110 used whole pixel alignments to export the photographic images of the part of finger successively.There is the part that repeats in the image of the part of the finger that is read respectively by sensor 110.Along with the translational speed of finger is slack-off, it is many that the ratio of the part of repetition becomes.
Picture signal from sensor 110 temporarily is kept in the storer 112, perhaps, is directly inputted in the unique point extraction unit 111, and the unique point of the image of a part of pointing is extracted out.Extract out for unique point, be described in addition.Unique point is stored in the storer 113 by the every pixel column arbitrarily of each frame.For example, first frame is written into memory area 1, the second frame and is written into memory area 2 ..., the n frame is written into memory area n.The unique point of each frame compares in unique point comparing section 114, calculates the translational speed of pointing on the position of the corresponding to pixel column of unique point in speed detection unit 116.In addition, simultaneously, the image of decision interframe is overlapping in overlapping detection unit 115, and promptly (correction) revised in the overlapping region of image in image correction portion (image synthesizes portion) 117, so that there be not the overlapping of image, shows in display part 118.This image is stored in the data preservation portion 120 when login, when authentication, compare with the data of in authentication department 119, being preserved, with the image of the finger that is corrected consistent, the inconsistent authentication result of indicator gauge first finger line together, under inconsistent situation, that points once more moves in display part 190, under corresponding to situation, enter the purpose (for example, the use of permission mobile phone makes and can converse) of this equipment.
For the unique point extracting method, following three kinds of methods are generally arranged: extract out from fingerprint image carina fingerprint characteristics such as structure (end points and take-off point) method (for example, the Marchal relations act), image (pattern) matching method of the similarity of evaluation map picture, laterally decomposing fingerprint image meticulously on (with the vertical direction of length direction of finger), the frequency analysis method that the deep or light difference of the carina of each image is carried out frequency analysis.
In addition, used the unique point of photographic images of a part of finger of a plurality of pixel alignments of a part to extract out although use identical input feature vector point extraction unit 111 here, extract out with the unique point of the photographic images of the part of the finger that uses whole pixel alignments, but also the unique point extraction unit can be set respectively, change the unique point extracting method, for example, unique point when using a plurality of pixel alignment of a part with image (pattern) matching method is extracted out, and the unique point when using whole pixel alignment with frequency analysis method is extracted out.
Figure 10 is the process flow diagram of the action of the above-mentioned image input of expression.So in addition action can be carried out corresponding program by generating, and is carried out by software.That is, at first, use a plurality of pixel alignments of the part of sensor 110 to come read output signal (step S1), judge to have or not finger (step S2) on the sensor 110 by the unique point that whether can extract fingerprint out.Under the situation of the unique point that can extract fingerprint out, carry out read (step S3) by the fingerprint image that shift action carried out of the finger that uses whole pixel alignments.On the other hand, under the situation of the unique point that can not extract fingerprint out, a plurality of pixel alignments that reuse the part of sensor 110 are come read output signal.
The image of the part of the finger that is read respectively by sensor 110, the unique point of the image of a part of pointing is extracted (step S4) out.Unique point is kept at (step S5) in the storer 113 by each pixel column in each frame.The unique point (step S6) that compares each frame, under the corresponding to situation of unique point a predetermined level is exceeded, calculate the translational speed of finger on the position of the corresponding to pixel column of unique point, in addition simultaneously, the overlapping of image of decision interframe is the overlapping region (step S7) of image.Then, revise (correction), so that there be not overlapping (the step S8) of image.Under the inconsistent situation of unique point, utilize the reading of fingerprint image of the shift action of the finger that has used whole pixel alignments once more.
Figure 11 be expression relatively the word that is identified as simple " M " shape of shape be subject as the time the synoptic diagram of view data.Data A represents the view data of the subject picture of t1 sometime, and data B represents that (view data of＞subject picture on t1), data C represents the view data of lap for t2 constantly.By the view data of comparing data A and the view data of data B, the data C of the lap of recognition image is modified to the view data of the subject picture of " M " shape.
Shown in the synoptic diagram of Figure 11 like that, under the situation that is identified as simple subject picture of shape relatively, the method for the feature of data A and data B as a comparison is preferably the images match method of the similarity of evaluation map picture.Under the situation that resembles the such more complicated of fingerprint, if relatively more a part of fingerprint, then the processing of frequency analysis method is fairly simple.The Marchal relations act, for the end points or the take-off point that detect carina, sensor area is preferably bigger.
Figure 12 and Figure 13 are the figure of a configuration example of expression sensor 110, and Figure 12 is the figure of the circuit block of expression sensor, and Figure 13 is the figure of expression sensor pixel portion.
In Figure 12, the 130th, the photoelectric conversion department of reading images, the 131st, the vertical scanning circuit (VSR) of the scan line of selection photoelectric conversion department 130, the 132nd, the temporary transient temporary transient preservation storer of preserving from the signal of photoelectric conversion department 130, the 133rd, the signal transmission from temporary transient preservation storer 132 is controlled to the differential amplifier of back level and the horizontal scanning circuit (HSR) of gain variable amplifier (PGA) 134.Signal from PGA 134 is transformed to digital signal by analog digital converter (A/D transducer) 136, imports the unique point extraction unit 111 shown in Figure 7 of back level.Vertical scanning circuit 131, the control signal of temporarily preserving storer 132, horizontal scanning circuit 133, PGA1 34, A/D transducer 136 origin self-timing generators (TG) 135 are controlled.This timing generator 135 is accepted the signal exported from switching part shown in Figure 7 121, to 131 outputs of vertical scanning circuit utilization has been used whole pixel alignments shown in Figure 8 finger shift action fingerprint image read the signal that begins to control.
Then, use Figure 13 that photoelectric conversion department shown in Figure 12 130 and temporary transient formation of preserving storer 132 are described.Photoelectric conversion department is constituted with the pixel amplifier form that is called as cmos sensor as used herein.Each pixel is formed by 6 elements.That is, pixel by photodiode PD, remove the reset switch M of the residual charge of photodiode PD
V, be used for the signal charge of photodiode PD is sent to the transmission switch M of the amplifier of back level
TX, be used to remove the reset switch M of residual charge of the gate portion of amplifier (buffer transistor)
RES, become the buffer transistor M of amplifier of the signal voltage of buffered gate portion
SF, buffer transistor M
SFSignal output control to switch M on the vertical signal line
SELConstitute.M
1And M
2Be the capacitor C that is used for send temporary transient preservation usefulness to from the signal of vertical signal line
S, C
NSwitch.M
3And M
4Be to be used for handle from capacitor C
S, C
NSignal S
1With noise N
1Send the switch of next differential amplifier (not shown) respectively to through the horizontal output line.All the blanket exposure device of pixel is equivalent to be used to sum up the switch M that resets
V, M
RESAnd the transmission switch M that is used to sum up transmission
TXBe to be reset to by transmitting switch M between the blanket exposure period of pixel all from whole pixels
TXDuring till the signal that is produced transmits and finishes, by lighting LED 1, can access the fingerprint image of finger during this period.
The signal of photodiode PD is transmitted to capacitor C
S, the reset noise of pixel amplifier (buffer transistor MSF) is transmitted to capacitor C
N
Use Figure 14 that the action sequence of sensor is described below.
The basic driver sequential of Figure 14 is: after whole pixels of sensor were summed up removing comprehensively, the LED 1 shown in Fig. 8 (B) was lighted, and the part of fingerprint is read.This signal is by the blanket amplifier gate portion (gate portion of buffer transistor MSF) that sends each pixel to.The signal of gate portion is transmitted to capacitor C by every row
SPreserve.Then, gate portion is reset, and its noise is stored in capacitor C
NIn.Signal and noise are sent to the differential amplifier of back level simultaneously, remove denoising from signal.This action repeats in every row.
Use Figure 14 to describe above-mentioned action in detail.At first, when the finger on the sensor 110 was detected, during T1, the action launching of sensor was from timing generator 135 output control signal φ
StartAccept control signal φ
StartVertical scanning circuit 131 during T2 at first signal psi
C1, φ
C2 (n), φ
C2 (n+1)... as the H level, connect switch M
V, M
RES, gate portion (the buffer transistor M of removing photodiode PD and amplifier in whole pixels
SFGate portion).Then, make signal psi
PIn whole pixels,, connect and transmit switch M simultaneously as the H level
TX, at this moment, LED 1 is lighted (φ
LEDBe the H level), shine on the mobile finger, carry out light-to-current inversion from the reflected light of finger by the photodiode PD of each pixel, to gate portion (the buffer transistor M of amplifier
SFGate portion) send signal from photodiode PD.Fingerprint image carries out reading of 20 row.And, make transmission switch M
TXUnder the state that turn-offs, during in the T3, light LED 1, then, connect and transmit switch M
TX, and can transmit signal.Control so that during the lighting of LED 1 between (during the rayed) and the exposure period that is produced by the blanket exposure device of whole pixels consistent or be between this exposure period in, thus, can reduce power consumption.The output signal of amplifier is by making signal psi
Sel (n)And φ
SBecome the H level, and connect switch M
SELAnd M
1, thus, be stored in capacitor C
SIn.
Then, make signal psi
C2 (n)Become the H level, make switch M
RESConnect, make amplifier gate portion (buffer transistor M
SFGate portion) be reset.Then, make signal psi
Sel (n)Become the H level with φ N, and connect switch M
SELAnd M
2, thus, reset noise is transmitted to capacitor C
NAt this moment, photodiode PD in order to prevent because ambient light etc. and, make signal psi by the leakage of electric charge that charge excess produced to gate portion
C1Become the H level, switch M
VConnect, be controlled on the set potential.Be stored in capacitor C
S, C
NInterior signal and noise are sent to the back level successively and handle.Then, when the signal of the capable pixel of n and noise are read successively, the output signal of the amplifier of the capable pixel of the n+1 that follows is sent to capacitor C in every row
S, reset noise is sent to capacitor C
N, carry out the action of reading that transmits processing in the back level.Like this, carry out reading of each pixel column.T4 whole picture element signals read into during the level of back during this time.When on detecting sensor 110, not pointing, from timing generator 35 output signal φ
End, the action of stop sensor.
Figure 15 represents to judge the sequential chart of pointing the sequential example that has or not.Repeat to drive a plurality of pixel alignments of the part of sensor shown in Figure 9 110, obtain data A1, data A2, data A3 ... (each data representation is from the output data of a plurality of pixel alignments of a part).Then, by these data, when detecting the unique point that becomes finger for the state of not extracting (finger is placed) from the unique point of finger out and be drawn out of the state of (having placed finger), utilize the reading of fingerprint image of the shift action of the finger that has used whole pixel alignments.
At this, the establishing method of reading the time of whole pixels is important.That is, need the top speed that moves of anticipation or default finger, the frame speed of decision sensor.When frame speed was slow, because finger is mobile, fingerprint was fuzzy, perhaps, when frame speed is too fast, under the mobile slow situation of finger, must prepare a plurality of video memories.
In the present embodiment, it is square that Pixel Dimensions is set at 30 μ m, and the highest translational speed of finger is set at 1cm/0.1sec.At this moment, if during LED lights be 28 μ sec, the mobile of finger during this period is 2.8 μ m, and the fuzzy part of image or the sampling error of interframe are below 1/10 of max picture size.When establish image overlapping about 1.4msec carries out repetition in 150 μ m for after minimum 5 pixels, frame rate is about 714 frame/seconds.Because control signal φ
ReadBe about 1.4msec during this time, so signal output frequency is about 8MHz.In the time of during further shortening LED lights, can reduce the sampling error of image, improve the fingerprint repeatability after image synthesizes, so can improve authentication precision.
If it is synthetic to be with it that prerequisite driving sensor carries out image described later, the unique point extraction method also can be realized satisfiable authentication with frequency analysis method, images match method, relations act.But (authentication precision that needs) set unique point extracting method or frame speed and got final product as required.
In addition, method as fingerprint detection, except to the finger irradiates light, with its reflected light of light sensors or see through light (perhaps scattering/see through light) and detect outside the measure of optical profile type of fingerprint, also exist finger is placed on the semiconductor transducers such as piezoelectric element, electrostatic condenser detecting element, detector unit, detect the semiconductor transducer mode of fingerprint by detected pressures, electric capacity, temperature, can use any detection method in the present invention.But, in the present invention, because the part of live image photographic images repeatedly then requires the translational speed fast sensor of the reading speed of sensor 10 than finger.
Another embodiment to image-input device of the present invention describes below.Present embodiment is the beginning of moving the driving of controlling Figure 14 of judging finger.
Figure 16 is another system diagram of the image-input device of present embodiment.In sequential chart shown in Figure 15, can sense data judge moving of finger.In Figure 16,, use identical label for the component parts identical with Fig. 7.
As shown in Figure 9, use a plurality of pixel alignments of the part of sensor 110 to come read output signal, judge on the sensor 110 to have or not finger.Promptly, as shown in Figure 15, repeat to drive a plurality of pixel alignments of the part of sensor shown in Figure 9 110, obtain data A1, data A2, data A3 ... (each data representation is from the output data of a plurality of pixel alignments of a part), through storer 112 or direct input feature vector point extraction unit 111, the unique point of a part of image of pointing is extracted out these data.Unique point is stored in the storer 113 each pixel column in each data.For example, data A1 is written into memory area 1, and data A2 is written into memory area 2 ..., data Am is written into memory area m, and (m is for less than the number of n, m＜n).The unique point of each data compares in unique point comparing section 114, image data A1 and data A2, data A2 and the such comparing data of data A3, fiducial value in data surpasses under the inconsistent situation of certain setting value, unique point comparing section 114 detects finger and moves, in addition, on the position of the corresponding to pixel column of unique point, calculate the translational speed of finger by speed detection unit 116.Send switching signals from unique point comparing section 114 or unique point comparing section 114 and speed detection unit 116 to switching part 122, switching part 122 is to the signal of the beginning of the driving of sequential generating unit 136 output control Figure 14 of sensor 110.In addition, owing to, can be suitable for this translational speed and come driving sensor, so just passable with a spot of storer from the signal of speed detection unit 116 to the level of the translational speed of the sequential generating unit 136 transmission expression fingers of switching part 122 or sensor 110.And LED 101 lights the power consumption that is produced and also can reduce.And then, owing to judging under the situation about moving of finger, not the whole pixel alignment of scanning, so can seek the low consumption electrification of system.
Use Figure 11 that data overlapping of interframe of the driving of Figure 14 is described.At this, for the purpose of simplifying the description, fingerprint is become character M be used as subject, character M is moved on sensor.The data B's of data A among the frame A and frame B is overlapping, and data C is that the overlapping of obtaining of image is lap.If exist slow, the overlapping data that move of subject to increase, the elongated such problem of frame driving time still, is come driving sensor by the translational speed that cooperates subject, just can reduce overlapping data, and reduces memory span.
As described above, according to present embodiment, owing to use little sensor, then cost is low, can realize miniaturization.And, can reduce power consumption, high-precision authentication can be carried out, and portable equipment can be used for.
Below embodiment 3 is described.
Figure 18 A is the figure that moves that schematically represents sensor and finger, and Figure 18 B is the figure that briefly constitute of expression by the fingerprint reading device that optical profile type realized.In the present embodiment, the usage level pixel count is 540 pixels, and Vertical number of pixels is the sensor 210 of 20 pixels.For example, when carrying out finger print identifying, using Vertical number of pixels usually is about 500 pixels, and therefore, in contrast to this, the size sensor of Shi Yonging is very little in the present embodiment.And, because in the present embodiment, not to read all images by sensor, get final product but read its a part of image, so Vertical number of pixels also is to get final product reading the sufficient pixel count of a part of image.On sensor 210 (as required; guard blocks such as configuration sheet glass and fibre plate on sensor 210) place finger; in the present embodiment; shown in Figure 18 A like that; make finger on respect to the vertical direction of the length direction of sensor 210, move (representing), reading images successively with the arrow among Figure 18 A.
Shown in Figure 18 B like that, from the sidepiece of finger (can be go up direction, oblique upper to, down direction, tiltedly descend direction, but in order to seek the miniaturization of device, preferably from being from sidepiece) utilize LED 1 irradiates light that becomes light irradiation device, in finger, make light scattering/see through, this light incides the sensor 210 from finger, obtains fingerprint image.
Figure 20 be expression relatively the identification of shape be the word of simple " M " shape be subject as the time the synoptic diagram of view data.Data A represents the view data of the subject picture of t1 sometime, and data B represents that (view data of＞subject picture on t1), data C represents the view data of lap for t2 constantly.Under corresponding situation, if slow, the then overlapping data that move of subject increase, the frame driving time is elongated.Therefore, in the present embodiment, come driving sensor, can reduce overlapping data, reduce the number of times that whole pixels are read, thus, can reduce memory span by the translational speed that cooperates subject.
At Figure 19 A, represented the concept map of the action of present embodiment among the B.Shown in Figure 19 A like that, except using Figure 18 A, outside the playback mode (being called whole pixel playback modes) of the fingerprint image of whole pixel alignments of the sensor 210 of 18B explanation (is the alignment of 20 row at this), also be provided with the playback mode (being called the partial pixel playback mode) of fingerprint image of a plurality of pixel alignments of a part of using sensor 10, shown in Figure 19 B like that, at first, in whole pixel playback modes, by sensor 210 reading of data A, then, repeating part pixel playback mode, when at resulting data B, data C, ... in have with data A in the corresponding to data of data A ' time, obtain carrying out then the data D that reads by whole pixel playback modes.Utilize the action of present embodiment, the lap L2 that utilizes the data A of whole pixel playback modes and data D when repeating whole pixel playback mode (after data A reads, do not carry out the partial pixel playback mode, dot the data when carrying out whole pixel playback mode) lap L1 and comparison, can diminish.Like this, even under the slow situation of translational speed of finger, also can reduce lap.
Figure 17 is the system diagram of an embodiment of image-input device of the present invention.As reference Figure 19 A, when 19B describes, in the existence of having confirmed the finger on the sensor 210, point move after, sensor 210 uses whole pixel alignments to export the photographic images (the data A among Figure 19 B) of the part of finger.Picture signal from sensor 210 is temporarily stored in storer 212, perhaps, directly is transfused in the unique point extraction unit 211, and the unique point of a part of image of pointing is extracted out.Extracting relevant content out with unique point describes in addition.Unique point is pressed each pixel column arbitrarily as the unique point of first frame, is stored in the memory area 1 of storer 213.
Then, as Figure 19 A, shown in the 19B like that, a plurality of pixel alignments of the part of driving sensor 210 obtain photographic images (the data B of Figure 19 B), these data B carries out the unique point of image and extracts out through storer 212 or direct input feature vector point extraction unit 211.Unique point is stored in the memory area 2 of storer 213 by every pixel column.The unique point of data A ' in the data A that is stored in memory area 1 compares in unique point comparing section 214 with the unique point of the data B that is stored in memory area 2, consistent with the unique point of data B or surpass under the situation of certain setting value in the unique point of data A ', unique point comparing section 214 transmits control signal to switching part 222, and switching part 222 switches to sensor 210 under whole pixel playback modes.In addition, the translational speed of calculating at the locational finger of the corresponding to pixel column of unique point by speed detection unit 216.In addition, in storer 213, do not preserve the unique point of data B,, compare with the unique point of data A ' from unique point extraction unit 211 direct input feature vector point comparing sections 214.
If the unique point of the unique point of data A ' and data B is inconsistent, a plurality of pixel alignments of the part of driving sensor 210 obtain photographic images (the data C of Figure 19 B), identical with above-mentioned action, the unique point of the unique point of comparing data A ' and data C, carry out partial pixel and read, consistent with the unique point of data A ' or surpass till certain setting value up to the unique point of reading the data that obtain by partial pixel.
When switching to by switching part 222 under whole pixel playback modes, carry out whole pixels when reading, identical with above-mentioned action, picture signal from sensor 210 is temporarily stored in storer 212 as data D, perhaps, directly be transfused to unique point extraction unit 211, the unique point of the parts of images of pointing is extracted out.Unique point, is stored in the memory area 2 of storer 213 by each any pixel column as the unique point of second frame.In addition, carrying out whole pixels when reading, it is deleted to read the data B that is stored or the data C that are produced by partial pixel in memory area 2.
Like this, after all pixels are read, carry out partial pixel and read, a part of reading the data that produced when whole pixels and partial pixel are read the data that produced when consistent or roughly consistent, execution whole pixels of carrying out following action of reading successively.Therefore, can detect the translational speed of finger, set frame speed sensor.In the case,, also can correspondingly with it change frame speed, adjust overlapping (can change frame speed) corresponding to the variation of translational speed even the translational speed of finger changes.In addition, under the situation that the translational speed of pointing less changes, set speed (frame speed) at first, also can after this be set, not carry out partial pixel and read, after whole pixels of following are read, move with this frame speed from data A to data D.
Like this, the unique point of the less data of lap is stored in the storer 213 successively.For example, read first frame that is produced by whole pixels and be written into memory area 1, the second frame and be written into memory area 2 ..., the n frame is written into memory area n.The unique point of every frame is compared by unique point comparing section 214, on the position of the corresponding to pixel column of unique point, is calculated the translational speed of finger by speed detection unit 216.In addition, simultaneously, be the overlapping region of image by the overlapping of image of overlapping detection unit 15 decision interframe, revise (correction) and synthesize by image correction portion (image synthesize portion) 217, so that do not have the overlapping of image, and on display part 218, show.This image is stored in the data preservation portion 220 when login, when authentication, compare with the data of in authentication department 219, being preserved, with the image of the finger that is corrected consistent, the inconsistent authentication result of indicator gauge first finger line together, under inconsistent situation, that points once more moves on display part 218, under corresponding to situation, enter the purpose (for example, the use of permission mobile phone allows to converse) of this equipment.
For the unique point extracting method, be generally following three kinds of methods: extract out from fingerprint image carina fingerprint characteristics such as structure (end points and take-off point) method (for example, the Marchal relations act), image (figure) matching method of the similarity of evaluation map picture, laterally decomposing fingerprint image meticulously on (with the vertical direction of length direction of finger), the frequency analysis method that the deep or light difference of the carina of each image is carried out frequency analysis.
Shown in the synoptic diagram of Figure 20 like that, be under the situation of simple subject picture in the identification of shape relatively, the method for the feature of data A and data B as a comparison is preferably the images match method of the similarity of evaluation map picture.Resemble under the situation of the such more complicated of fingerprint, if relatively more a part of fingerprint, then the processing of frequency analysis method is fairly simple.End points or the take-off point of Marchal relations act in order to detect carina, sensor area is preferably bigger.
Figure 21 and Figure 22 are the figure of a configuration example of expression sensor 210, and Figure 21 is the figure of the circuit block of expression sensor, and Figure 22 is the figure of expression sensor pixel portion.
In Figure 21, the 230th, the photoelectric conversion department of reading images, the 231st, the vertical scanning circuit (VSR) of the scan line of selection photoelectric conversion department 230, the 232nd, the temporary transient temporary transient preservation storer of preserving from the signal of photoelectric conversion department 230, the 233rd, the signal transmission from temporary transient preservation storer 32 is controlled to the differential amplifier of back level and the horizontal scanning circuit (HSR) of gain variable amplifier (PGA) 234.Signal from PGA 234 is transformed to digital signal by analog digital converter (A/D transducer) 236, imports the unique point extraction unit 211 shown in Figure 17 of back level.Vertical scanning circuit 231, the control signal of temporarily preserving storer 232, horizontal scanning circuit 233, PGA 234, A/D transducer 236 origin self-timing generators 235 are controlled.
Then, use Figure 22 that photoelectric conversion department shown in Figure 21 230 and temporary transient formation of preserving storer 232 are described.Photoelectric conversion department used herein is constituted with the pixel amplifier form that is called as cmos sensor.Each pixel is formed by 6 elements.That is, pixel by photodiode PD, remove the reset switch M of the residual charge of photodiode PD
V, be used for the signal charge of photodiode PD is sent to the transmission switch M of the amplifier of back level
TX, be used to remove the reset switch M of residual charge of the gate portion of amplifier (buffer transistor)
RES, become the buffer transistor M of amplifier of the signal voltage of buffered gate portion
SF, buffer transistor M
SFSignal output control to switch M on the vertical signal line
SELConstitute.M
1And M
2Be the capacitor C that is used for send temporary transient preservation usefulness to from the signal of vertical signal line
S, C
NSwitch.M
3And M
4Be to be used for handle from capacitor C
S, C
NSignal S
1With noise N
1Send the switch of next differential amplifier (not shown) respectively to through the horizontal output line.All the blanket exposure device of pixel is equivalent to be used to sum up the switch M that resets
V, M
RESAnd the transmission switch M that is used to sum up transmission
TXBe to be reset to by transmitting switch M between the blanket exposure period of pixel all from whole pixels
TXDuring till the signal that produced transmits and finishes, by lighting LED 1, and can access the fingerprint image of finger during this period.
The signal of photodiode PD is transmitted to capacitor C
S, the reset noise of pixel amplifier (buffer transistor MSF) is transmitted to capacitor C
N
Use Figure 23 to come the action sequence of sensor is described below.
The basic driver sequential of Figure 23 is: after whole pixels of sensor were summed up removing comprehensively, the LED 201 shown in Figure 18 (B) was lighted, and the part of fingerprint is read.This signal is by the blanket amplifier gate portion (gate portion of buffer transistor MSF) that sends each pixel to.The signal of gate portion is transmitted to capacitor C by every row
SPreserve.Then, gate portion is reset, and its noise is stored in capacitor C
NIn.Signal and noise are sent to the differential amplifier of back level simultaneously, remove denoising from signal.This action repeats every row.
Use Figure 23 to illustrate in greater detail above-mentioned action.At first, when the finger on the sensor 210 was detected, during T1, the action launching of sensor was from timing generator 235 output control signal φ
StartAccept control signal φ
StartVertical scanning circuit 231 during T2 at first signal psi
C1, φ
C2 (n), φ
C2 (n+1)... as the H level, connect switch M
V, M
RES, the gate portion (gate portion of buffer transistor MSF) of removing photodiode PD and amplifier in whole pixels.Then, make signal psi
PIn whole pixels,, connect and transmit switch M simultaneously as the H level
TX, at this moment, LED 201 is lighted (φ
LEDBe the H level), shine on the mobile finger, carry out light-to-current inversion from the reflected light of finger by the photodiode PD of each pixel, send signal to the gate portion (gate portion of buffer transistor MSF) of amplifier from photodiode PD.Fingerprint image carries out reading of 20 row.And, make transmission switch M
TXUnder the state that turn-offs, during in the T3, light LED 201, then, connect and transmit switch M
TX, and can transmit signal.Control so that during the lighting of LED 201 between (during the rayed) and the exposure period that is produced by the blanket exposure device of whole pixels consistent or be between this exposure period in, thus, can reduce power consumption.The output signal of amplifier is by making signal psi
Sel (n)And φ
SBecome the H level, and connect switch M
SELAnd M
1, be stored in capacitor C
SIn.
Then, make signal psi
C2 (n)Become the H level, make switch M
RESConnect, amplifier gate portion (gate portion of buffer transistor MSF) is reset.Then, make signal psi
Sel (n)And φ
NBecome the H level, and connect switch M
SELAnd M
2, thus, reset noise is transmitted to capacitor C
NAt this moment, photodiode PD in order to prevent because ambient light etc. and, make signal psi by the leakage of electric charge that charge excess produced to gate portion
C1Become the H level, switch M
VConnect, be controlled on the set potential.Be stored in capacitor C
S, C
NInterior signal and noise are sent to the back level successively and handle.Then, when the signal of the capable pixel of n and noise are read successively, the output signal of the amplifier of the capable pixel of the n+1 that follows is sent to capacitor C in every row
S, reset noise is sent to capacitor C
N, carry out the action of reading that transmits processing in the back level.Like this, carry out reading of each pixel column.T4 whole picture element signals read into during the level of back during this time.When on detecting sensor 210, not pointing, from timing generator 35 output signal φ
End, the action of stop sensor.
The sequential chart of the action shown in Figure 24 presentation graphs 19B.In whole pixel playback modes, by sensor 210 reading of data A, then, repeating part pixel playback mode, when resulting data B, data C ... in have with data A in the corresponding to data of data A ' time, obtain carrying out then the data D that reads by whole pixel playback modes.By carrying out such action successively, all for finger, all carrying out reading of fingerprint under the pixel playback mode.
At this, the establishing method of reading the time of whole pixels is important.That is, need the top speed that moves of anticipation or default finger, the frame speed of decision sensor.When frame speed was slow, because finger is mobile, fingerprint was fuzzy, perhaps, when frame speed is too fast, under the mobile slow situation of finger, must prepare a plurality of video memories.
In the present embodiment, it is square that Pixel Dimensions is set at 30 μ m, and the highest translational speed of finger is set at 1cm/0.1sec.At this moment, if during LED lights be 28 μ sec, the mobile of finger during this period is 2.8 μ m, and the fuzzy part of image or the sampling error of interframe are below 1/10 of max picture size.Overlapping during for minimum 5 pixels when what establish image, about 1.4msec carries out repetition in 150 μ m, and frame rate is about 714 frame/seconds.Control signal φ
ReadBe about 1.4msec during this time, therefore, signal output frequency is about 8MHz.In the time of during further shortening LED lights, can reduce the sampling error of image, improve the fingerprint repeatability after image synthesizes, therefore, can improve authentication precision.
If with it is prerequisite, it is synthetic that driving sensor carries out image described later, and the unique point extraction method also can be realized satisfiable authentication with frequency analysis method, images match method, relations act.But (authentication precision that needs) set unique point extracting method or frame speed and got final product as required.
In addition, method as fingerprint detection, except to the finger irradiates light, with its reflected light of light sensors or see through light (perhaps scattering sees through light) and detect outside the measure of optical profile type of fingerprint, also exist finger is placed on the semiconductor transducers such as piezoelectric element, electrostatic condenser detecting element, detector unit, detect the semiconductor transducer mode of fingerprint by detected pressures, electric capacity, temperature, can use any detection method in the present invention.But, in the present invention, because the part of live image photographic images repeatedly then requires the translational speed fast sensor of the reading speed of sensor 10 than finger.
Then another embodiment to image-input device of the present invention describes.Present embodiment does not carry out partial pixel to be read, and reads mutual comparison by whole pixels, reduces the lap of data.
The concept map of in Figure 25, having represented the action of present embodiment.As shown in Figure 25, at first, by sensor 210 with primary whole pixel playback mode reading of data A, then, by sensor 210 with secondary whole pixel playback mode reading of data B.Data B ' in the data B in the memory area that is stored in storer 213 2 that in unique point comparing section 214, relatively under secondary whole pixel playback modes, obtains and the data A ' in the data A in the memory area 1 of being stored in that under for the first time whole pixel playback modes, obtains, when consistent or roughly consistent, as shown in Figure 25, then, by sensor 210 at whole reading of data C under the pixel playback modes for the third time.At this moment, data B is deleted from storer 213.In addition, when consistent or roughly consistent, use is resulting data B under secondary whole pixel playback modes, resulting data A carries out unique point relatively in unique point comparing section 214 and under primary whole pixel playback modes, carries out the authentication of fingerprint with action same as the previously described embodiments.
In unique point comparing section 214, relatively data C ' in the data C in the resultant memory area 2 that is stored in storer 213 and the data A ' in the data A in the memory area 1 of being stored in that under for the first time whole pixel playback modes, obtains under whole pixel playback modes for the third time, when consistent or roughly consistent, become the state of storage data C in the memory area 2 of storer 213, carry out whole pixels then and read.This illustrated the data comparison carried out in all pixels are read up to read point all finger print datas till.The data A of the whole pixel playback modes of the utilization that action produced of present embodiment and the lap L2 of data C are with all the data A of pixel playback mode and the lap L1 of data C compare little, by not using data B, even under the slow situation of translational speed of finger, also can reduce lap.Other action is identical with the action of the foregoing description, and in addition, in order to improve processing speed, data A ' is preferably near data A last, and the data B ' that compares, data C ' are preferably near the beginning of data B, data C.
The sequential chart of in Figure 26, having represented action shown in Figure 25.Under whole pixel playback modes by sensor 210 reading of data A, under whole pixel playback modes after this resulting data B, data C ... interior data B ', data C ' more whether with data A in data A ' consistent or roughly consistent (at this, data C ' is consistent with data A '), use has unanimity or roughly consistent data C, does not use the data B with inconsistent data B '.Like this, can reduce lap.
Figure 27 be in by the embodiment of Figure 19 B explanation expression the identification of shape relatively be the word of simple " M " shape be replaced into subject as the time the synoptic diagram of view data.Read the driving reading of data A that is carried out by whole pixels, read the driving reading of data B that is carried out by the partial pixel of following, from data A and the corresponding to locations of pixels of data B, judge the translational speed of subject, the driving that can utilize whole pixels then to read.
In addition, as the data of reading the driving of being carried out by whole pixels, can only utilize the data of the part among the data A.Figure 28 be expression the identification of shape relatively be the word of simple " M " shape be replaced into subject as the time the synoptic diagram of view data.If the data of utilizing the part among the data A are as the data of being read the driving of being carried out by whole pixels, read in the driving of being carried out at the partial pixel of following, use data B1, B2, B3, B4 among the data B, carry out the whether consistent or roughly consistent judgement of data A and data B.In the case, because the driving of following is the part for the read pixel alignment, so have the advantage that to judge fast.
Because without departing from the spirit and scope of the present invention, can realize the present invention's form of implementation in a large number and widely, be construed as the present invention and be not limited to specific form of implementation, but by additional claim defined.
Claims (1)
1. image-input device, by subject and the sensor that comprises a plurality of pixels are relatively moved, divide the image of above-mentioned subject and obtain a plurality of parts of images respectively successively as the part of the image of above-mentioned subject, so that have overlapping between resulting above-mentioned a plurality of parts of images, it is characterized in that, comprising:
Control circuit, according to the comparison of a plurality of above-mentioned parts of images data, control obtains the timing of a plurality of above-mentioned parts of images data that will synthesize; With
Image synthesizer is proofreaied and correct the lap of above-mentioned a plurality of parts of images of exporting from the sensor and to carry out image synthetic,
Wherein, above-mentioned control circuit comprises commutation circuit and testing circuit,
Described commutation circuit, switching first is read mode and second and is read mode, described first reads mode utilizes a plurality of pixels of the sensor to carry out image to read, described second reads mode utilizes carrying out image and read than reading a plurality of pixels that the pixel count that uses in the mode lacks above-mentioned first of the sensor
Described testing circuit, the subject for relatively moving with the sensor relatively reads mode with above-mentioned first and reads resulting parts of images data and after this read resulting parts of images data with the above-mentioned second reading mode of getting,
Reading resulting parts of images data with the above-mentioned second reading mode of getting and reading mode with above-mentioned first and read when between the resulting parts of images data corresponding to data being arranged when above-mentioned testing circuit detects, by above-mentioned commutation circuit, read mode from above-mentioned second and switch to above-mentioned first and read mode.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP043238/2002 | 2002-02-20 | ||
| JP2002043238A JP2003242489A (en) | 2002-02-20 | 2002-02-20 | Image input device and fingerprint recognizing device |
| JP046475/2002 | 2002-02-22 | ||
| JP046473/2002 | 2002-02-22 |
Related Parent Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB031045715A Division CN1235172C (en) | 2002-02-20 | 2003-02-18 | Image input device |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2006100959752A Division CN100430956C (en) | 2002-02-20 | 2003-02-18 | Image input apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1758267A CN1758267A (en) | 2006-04-12 |
| CN100362526C true CN100362526C (en) | 2008-01-16 |
Family
ID=27783089
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101170231A Expired - Fee Related CN100362526C (en) | 2002-02-20 | 2003-02-18 | Image input apparatus |
| CNB2006100959752A Expired - Fee Related CN100430956C (en) | 2002-02-20 | 2003-02-18 | Image input apparatus |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2006100959752A Expired - Fee Related CN100430956C (en) | 2002-02-20 | 2003-02-18 | Image input apparatus |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP2003242489A (en) |
| CN (2) | CN100362526C (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE60238281D1 (en) | 2002-09-17 | 2010-12-23 | Fujitsu Ltd | APPARATUS FOR RECORDING BIOLOGICAL INFORMATION AND BIOLOGICAL INFORMATION USING THE AUTHORIZATION DEVICE |
| JP3924558B2 (en) | 2003-11-17 | 2007-06-06 | 富士通株式会社 | Biological information collection device |
| JP4462988B2 (en) | 2004-04-13 | 2010-05-12 | Ｎｅｃインフロンティア株式会社 | Fingerprint reading method and fingerprint reading system |
| JP4411152B2 (en) | 2004-07-05 | 2010-02-10 | Ｎｅｃインフロンティア株式会社 | Fingerprint reading method, fingerprint reading system and program |
| JP4427039B2 (en) * | 2006-08-18 | 2010-03-03 | 富士通株式会社 | Biometric information acquisition apparatus and biometric information authentication apparatus |
| EP2455761B1 (en) * | 2010-11-23 | 2018-08-08 | Hach Lange GmbH | Method for detecting an optical marking on a laboratory analysis vessel |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1172308A (en) * | 1996-06-14 | 1998-02-04 | 汤姆森-无线电报总公司 | Fingerprint-reading system |
| WO2000046743A1 (en) * | 1999-02-01 | 2000-08-10 | Beonic Corporation Pty Ltd | Object recognition and tracking system |
| JP2002042136A (en) * | 2000-07-28 | 2002-02-08 | Nec Corp | Fingerprint identifying method and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6259108B1 (en) * | 1998-10-09 | 2001-07-10 | Kinetic Sciences Inc. | Fingerprint image optical input apparatus |
-
2002
- 2002-02-20 JP JP2002043238A patent/JP2003242489A/en active Pending
-
2003
- 2003-02-18 CN CNB2005101170231A patent/CN100362526C/en not_active Expired - Fee Related
- 2003-02-18 CN CNB2006100959752A patent/CN100430956C/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1172308A (en) * | 1996-06-14 | 1998-02-04 | 汤姆森-无线电报总公司 | Fingerprint-reading system |
| JPH1091769A (en) * | 1996-06-14 | 1998-04-10 | Thomson Csf | Fingerprint reading system |
| US6289114B1 (en) * | 1996-06-14 | 2001-09-11 | Thomson-Csf | Fingerprint-reading system |
| WO2000046743A1 (en) * | 1999-02-01 | 2000-08-10 | Beonic Corporation Pty Ltd | Object recognition and tracking system |
| JP2002042136A (en) * | 2000-07-28 | 2002-02-08 | Nec Corp | Fingerprint identifying method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100430956C (en) | 2008-11-05 |
| JP2003242489A (en) | 2003-08-29 |
| CN1873655A (en) | 2006-12-06 |
| CN1758267A (en) | 2006-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1442824A (en) | Image input device | |
| EP1318473B1 (en) | Image input apparatus and method correcting the illumination of the imaged object | |
| CN108121472B (en) | Touch display device, display panel, touch sensing method and circuit and driving circuit | |
| JP2005049193A (en) | Input device, electronic instrument, and method for driving input device | |
| US8254715B2 (en) | Method and apparatus of detecting image objects | |
| CN100362526C (en) | Image input apparatus | |
| US20100208308A1 (en) | Method and apparatus for recognizing characters | |
| CN111027516A (en) | Biological characteristic image acquisition device and method and intelligent equipment | |
| JP2886575B2 (en) | Fingerprint recognition method and device | |
| JP2005063020A (en) | Recognition device | |
| JP2004213689A (en) | Image input device and fingerprint recognition device | |
| TWI720660B (en) | Optical image residual image elimination method and electronic device using the same | |
| JP2003248818A (en) | Image input device and fingerprint recognition device | |
| TWI568237B (en) | Image capture method and image capture and synthesis method | |
| KR100700927B1 (en) | Method of scanning an image using surface coordinate values and device using thereof | |
| EP0950977A4 (en) | User identification method and unit | |
| SU732926A1 (en) | Device for symbol identification | |
| KR101712435B1 (en) | Regularization apparatus for depth information on edge | |
| SU1084835A1 (en) | Device for reading graphic information | |
| JPS5668869A (en) | Optical character reader | |
| JP2003248817A (en) | Image input device and fingerprint recognition device | |
| JPS6280779A (en) | Character reader | |
| JPS6015993B2 (en) | Manual scanning and reading of characters | |
| JPH03142580A (en) | Compact data terminal equipment with image input function | |
| JPH0816637A (en) | Circuit diagram with two-dimensional data code and waveform measurement system using the diagram |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080116 Termination date: 20150218 |
|
| EXPY | Termination of patent right or utility model |