CN202748806U - Image gray level information-based finger detection realization architecture - Google Patents
Image gray level information-based finger detection realization architecture Download PDFInfo
- Publication number
- CN202748806U CN202748806U CN201220477535.4U CN201220477535U CN202748806U CN 202748806 U CN202748806 U CN 202748806U CN 201220477535 U CN201220477535 U CN 201220477535U CN 202748806 U CN202748806 U CN 202748806U
- Authority
- CN
- China
- Prior art keywords
- finger
- value
- image
- threshold point
- adjacent
- 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 - Lifetime
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
Provided by the utility model is an image gray level information-based finger detection realization architecture that comprises a histogram statistics module, a threshold point generation module, a local difference maximum value calculation module for two adjacent images, an offset result calculation module for two adjacent images, and a finger detection module. The working process of all the modules are as follows: S1, generating a fingerprint image threshold point by the threshold point generation module according to the histogram arrangement; S2, calculating a local difference maximum value of two adjacent images and determining whether there is a certain difference between the two adjacent images by the local difference maximum value calculation module for two adjacent images; S3, determining whether a finger is taken away or is placed according to an image non-background distribution situation and the difference of the two adjacent images by the finger detection module; and S4, calculating the offset of the two adjacent images and determining whether the finger is moved or not during the image collection process by the offset result calculation module for two adjacent images. Compared with the existing finger detection method, the method employed by the utility model enables the finger detection realization architecture to be flexible.
Description
Technical field
The utility model relates to digital picture identification field, relates in particular to a kind of the finger based on gradation of image information and detects the realization framework.
Background technology
Existing finger detection technique detects finger by differential detection circuit and puts with finger and leave, and it detects principle mainly is by at sensor surface or innerly increase an electric parameter detection means, thermistor for example, voltage dependent resistor (VDR), the devices such as electric capacity.The temperature of electric capacity, inductance or the finger that forms according to sensor surface and finger itself, the physical parameters such as the pressure of surface of contact, the humidity electric parameters such as causing its voltage by the generation of electric parameter detection means or electric current that changes is changed, point and put, point and leave detection.Because the finger testing circuit places sensor surface or sensor internal, has increased like this cost and the volume of sensor.
Detect the variation of electric parameter by difference channel and point the easy error detection of detection.The difference channel finger detects based on electric principle of induction, when finger is placed on sensor surface, cause the physical parameters such as near the humidity that forms of sensor, pressure, temperature, electric field to change, thereby making detection means detect electric parameter changes, the size that difference finger testing circuit changes by detecting electric parameter has judged whether to point and has put.Detect the variable quantity of electric parameter and realize that with reference to electric threshold value finger detects by finger relatively, if finger detects the electric parameter variable quantity greater than the electric threshold value of predefined reference, thinking has finger to exist, otherwise thinks that finger is not placed on the sensor.
Because difference detecting method is based on the size of electric parameter variation and points the detection judgement, during object proximity sensor that any one physical parameter can be detected by the electric parameter detection means, can cause near the physical parameter of sensor to change, thereby cause the electric parameter of sensor to change.If the variable quantity of electric parameter is greater than with reference to electric threshold value, thereby think that detecting finger puts by mistake; When having bone-dry finger to be placed on the sensor surface, its electric parameter changes smaller, if electric parameter changes less than the electric threshold value of reference, thereby the error detection finger is not put.
Summary of the invention
In order to address the above problem, the utility model provides a kind of the finger based on gradation of image information to detect the realization framework, it is characterized in that, comprising:
The statistics with histogram module: according to the gray-scale value of each pixel of scan image, the histogram of real-time statistics present frame fingerprint image is used for generating threshold point;
Threshold point generation module: link to each other with the statistics with histogram module, calculate in real time upper limit threshold point H, middle limit threshold point M, the lower threshold point L of present frame and calculate the higher limit C1 of this segment, the statistical threshold SUM2 between the second Statistical Area according to the statistical threshold SUM1 between the first Statistical Area according to histogram distribution and calculate the higher limit C2 of this segment, thereby obtain the non-background information distribution situation of image;
Partial error's opposite sex maximal value of adjacent two two field pictures is asked for module: be used for partial error's opposite sex value of every kind of overlay area of calculating two two field pictures, and ask for partial error's opposite sex maximal value under all coverage conditions, judge whether adjacent two two field pictures there are differences;
The side-play amount result of adjacent two two field pictures asks for module: ask for partial error's opposite sex value of every kind of overlay area that module obtains according to adjacent two two field picture partial errors opposite sex maximal value, find out the different in nature minimum value of partial error under all coverage conditions; The horizontal relative displacement of the central point of two two field pictures is horizontal offset under the coverage condition of partial error's opposite sex minimum value, and vertically opposite displacement is vertical offset; When two frames cover fully and partial error's opposite sex hour, the relative displacement of the central point when preferentially selecting two frames to cover fully is as side-play amount;
Finger detection module: ask for module with side-play amount result that the partial error of threshold point generation module, adjacent two two field pictures opposite sex maximal value is asked for module, adjacent two two field pictures and link to each other, judge the quantity of the non-background half-tone information of image according to threshold point, get rid of the sensor sheet face portion dirty by the situation of error detection; According to the partial error of adjacent two two field pictures opposite sex maximal value, judge the whether property of there are differences of two two field pictures, get rid of sensor surface all dirty space-time collection by the situation of error detection; According to the side-play amount of adjacent two two field pictures, judge whether finger moves.
Described threshold point generation module is according to the total number VH of pixel in the available gray-scale threshold value G_TH statistics available gray-scale interval, with reference to formula (2)
Wherein x is the gradation of image value, and hist (x) is the histogram of x for gray-scale value, and effective gamma G_TH is less than or equal to the gradation of image scope, and G_TH is according to gradation of image scope capable of regulating;
Described threshold point generation module scans the histogram hist (x) of gray-scale value from 0 to K successively, and every run-down histogram carries out integration to histogram, obtains
Described K is the gradation of image maximal value, and i is gray-scale value corresponding to the current histogram that is scanned, 0≤i≤K.
Described threshold point generation module obtains threshold parameter according to the total number VH of the pixel in the available gray-scale interval, lower threshold point scale-up factor LR, middle limit threshold point scale-up factor MR, upper limit threshold point scale-up factor HR, available gray-scale threshold value G_TH according to predefined threshold parameter selection rule.
Described predefined threshold parameter selection rule is:
Work as y
I-1≤ VH * LR and y
iDuring＞VH * LR, the value of i is exactly histogram lower threshold parameter L, if L＞G_TH, then L=G_TH; 0＜LR＜1;
Work as y
I-1≤ VH * MR and y
iDuring＞VH * MR, the value of i is exactly limit threshold parameter M in the histogram, if M＞G_TH, then M=G_TH; 0＜MR＜1 and LR＜MR;
Work as y
I-1≤ VH * HR and y
iDuring＞VH * HR, the value of i is exactly histogram upper limit threshold Parameter H, if H＞G_TH, then H=G_TH; 0＜HR＜1 and MR＜HR;
Work as y
I-1≤ SUM1 and y
iDuring＞SUM1, the i value of this moment is exactly C1, if during i=K and y
iDuring≤SUM1, C1 equals K;
Work as y
I-1≤ SUM2 and y
iDuring＞SUM2, the i value of this moment is exactly C2, if during i=K and y
iDuring≤SUM2, C2 equals K;
SUM1＞SUM2。
The partial error of described adjacent two two field pictures opposite sex maximal value is asked for partial data or the total data that module chooses the overlay area in and is calculated the different in nature value of partial error, referring to formula (3), and from partial error's opposite sex value, find out partial error's opposite sex maximal value under all coverage conditions;
dif(u,v)＝∑(slice1(m,n)-slice2(m+u,n+v))
2 （3）
Slice1 represents the previous frame image in the formula (3), slice2 represents current frame image, m, horizontal coordinate and the vertical coordinate of n presentation video pixel, and m, the scope of n is the scope that participates in the pixel of computing in the selected overlay area, and u, v are horizontal relative displacement and the vertically opposite displacement of the central point of adjacent two two field pictures.
The finger of described finger detection module is put and is left basis for estimation with finger and be:
A, judge the non-background information amount of image that lower threshold point is following according to the threshold point distribution situation, if lower threshold is o'clock more than or equal to the higher limit between the first Statistical Area, the non-background number of image below the expression lower threshold point is more than or equal to statistical threshold between the first gray area, and expression detects partially wet finger and puts;
B, judge according to the threshold point distribution situation whether the non-background half-tone information of image mainly is distributed in more than the upper limit threshold point, if higher limit is greater than the upper limit threshold point between the second Statistical Area, the number below the expression upper limit threshold point is less than threshold value between the second Statistical Area, and expression detects finger and leaves;
C, whether there is some difference to judge adjacent two two field pictures according to partial error's opposite sex maximal value of the non-background information amount of image below the limit threshold point in the judgement of threshold point distribution situation and adjacent two two field pictures, if middle limit threshold point is put threshold value more than or equal to partial error's opposite sex maximal value of the higher limit between the first Statistical Area and adjacent two two field pictures greater than finger, expression detects finger and puts;
D, judge according to the threshold point distribution situation whether the non-background information of image concentrates on more than the middle limit threshold point, and judge the otherness that adjacent two two field pictures exist according to the partial error of adjacent two two field pictures opposite sex maximal value, if the higher limit between the second Statistical Area is left threshold value greater than partial error's opposite sex maximal value of middle limit threshold point and adjacent two two field pictures less than finger, judge that detecting finger leaves.
Described finger detection module to pointing mobile basis for estimation is: put in the situation that detect finger, judge whether one of them is not 0 for the horizontal offset of present frame and previous frame image and vertical offset, if one of them is not 0, it is mobile that expression detects finger, do not put if detect finger, it is mobile that expression does not detect finger.
The threshold point that the utility model is tried to achieve according to histogram distribution, partial error's opposite sex maximal value of adjacent two two field pictures and the side-play amount of adjacent two two field pictures are pointed to put with finger and are left detection, can get rid of dirty background by error detection, dried finger is missed the situation of survey.With respect to existing detection technique, the utility model cost is lower, and power consumption is lower, takies sensor area little.Finger detection system of the present utility model has increased the mobile sign that detects of finger, and this finger movable signal can carry out Mobile Telephone Gps in conjunction with its up and down coordinate that moves and use the data splicing that the scratch type sensor gathers etc.
Description of drawings
Fig. 1 is that sensor surface is clean and be placed on image and the histogram thereof that gathers on the sensor without finger;
Fig. 2 is image and the histogram thereof that the sensor sheet face portion is dirty and gather when putting without finger;
Fig. 3 is whole dirty and the image and the histogram thereof that gather when putting without finger of sensor surface;
Image and corresponding histogram thereof that Fig. 4 gathers on the sensor for wet finger is put in;
Image and corresponding histogram thereof that Fig. 5 gathers on the sensor for normal finger is put in;
Fig. 6 is put in image and the corresponding histogram thereof that gathers on the sensor for doing finger;
Fig. 7 is move up and down rule and the overlay area of adjacent two two field pictures;
Fig. 8 is horizontal relative displacement and the vertically opposite displacement of the central point of adjacent two two field pictures;
Fig. 9 is finger detection system block diagram;
Figure 10 is the finger testing result figure of the utility model continuous acquisition fingerprint image, and wherein Figure 10-1 is the fingerprint image that wet finger collection arrives, Figure 10-2 fingerprint image that arrives for normal finger collection, Figure 10-3 fingerprint image that arrives for dried finger collection.
Embodiment
Below in conjunction with accompanying drawing preferred embodiment of the present utility model is described, specific embodiment described herein is only in order to explaining the utility model, and is not used in restriction the utility model.
Referring to Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, a complete finger detection system considers that mainly following four kinds of situations error detection: ⅰ can not occur pointing, sensor surface is cleaner, finger is not placed on when gathering on the sensor, can not detect finger and be placed on the sensor.As shown in Figure 1, when sensor surface is clean, being placed on the gradation of image information that collects on the sensor without finger is image background information, because the image background value is very large, the threshold point that calculates is very large, the image of continuous acquisition two frames all is image background information, and the partial error's opposite sex maximal value that calculates adjacent two two field pictures is very little, and the situation of error detection can not appear in this situation.
ⅱ, sensor surface are dirty, and finger is not placed on when gathering on the sensor, can not error detection be placed on the sensor to finger.Such as Fig. 2, shown in Figure 3.When the whole surface of sensor is all dirty, carry out the non-background information of image that sky collects many, very little by the partial error's opposite sex maximal value that calculates adjacent two two field pictures; When the sensor sheet face portion is dirty, carry out the non-background information of image that sky collects fewer.The utility model by judging adjacent two two field pictures partial error opposite sex maximal value and the non-background information of image of present frame how much get rid of dirty background by the situation of error detection.
ⅲ, when there being partially wet finger to be placed on the sensor, can detect accurately finger and put, point and leave; As shown in Figure 4, when having partially wet finger to be placed on the sensor, the non-background information of the image of collection is very many, and namely mainly to be distributed near gray-scale value be 0 zone to the non-background information of image.
ⅳ, when having normal finger or partially dried finger to be placed on the sensor, can detect accurately the finger put, point and leave.Such as Fig. 5, shown in Figure 6, when normal finger or partially dried when being placed on the sensor is arranged, the non-background information of the image that collects is many, and the texture of adjacent two two field pictures has certain difference, and the partial error's opposite sex maximal value that calculates adjacent two two field pictures is enough large.
According to above principle, the utility model will be pointed to detect and realize that framework hangs over realization SOC(System On Chip SOC (system on a chip) on the system bus as an IP kernel), (can change the scale-up factor of asking threshold point by programmable parameter designing, statistical threshold between Statistical Area, finger is put threshold value, finger leaves threshold value etc.), total data or the partial data that can choose according to demand the overlay area calculate partial error's opposite sex value, with less resource, detect accurately to point to put with finger and leave, have very large dirigibility.As shown in Figure 9, a kind of finger based on gradation of image information detected the realization framework, comprising:
The statistics with histogram module: according to the gray-scale value of each pixel of scan image, the histogram of real-time statistics present frame fingerprint image is used for generating threshold point;
Threshold point generation module: link to each other with the statistics with histogram module, calculate in real time upper limit threshold point H, middle limit threshold point M, the lower threshold point L of present frame and calculate the higher limit C1 of this segment, the statistical threshold SUM2 between the second Statistical Area according to the statistical threshold SUM1 between the first Statistical Area according to histogram distribution and calculate the higher limit C2 of this segment, thereby obtain the non-background information distribution situation of image;
Partial error's opposite sex maximal value of adjacent two two field pictures is asked for module: be used for partial error's opposite sex value of every kind of overlay area of calculating two two field pictures, and ask for partial error's opposite sex maximal value under all coverage conditions, judge whether adjacent two two field pictures there are differences;
The side-play amount result of adjacent two two field pictures asks for module: ask for partial error's opposite sex value of every kind of overlay area that module obtains according to adjacent two two field picture partial errors opposite sex maximal value, find out the different in nature minimum value of partial error under all coverage conditions; The horizontal relative displacement of the central point of two two field pictures is horizontal offset under the coverage condition of partial error's opposite sex minimum value, and vertically opposite displacement is vertical offset; When two frames cover fully and partial error's opposite sex hour, the relative displacement of the central point when preferentially selecting two frames to cover fully is as side-play amount;
Finger detection module: ask for module with side-play amount result that the partial error of threshold point generation module, adjacent two two field pictures opposite sex maximal value is asked for module, adjacent two two field pictures and link to each other, judge the quantity of the non-background half-tone information of image according to threshold point, get rid of the sensor sheet face portion dirty by the situation of error detection; According to the partial error of adjacent two two field pictures opposite sex maximal value, judge the whether property of there are differences of two two field pictures, get rid of sensor surface all dirty space-time collection by the situation of error detection; According to the side-play amount of adjacent two two field pictures, judge whether finger moves.
Referring to Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, the horizontal ordinate that each stain of histogram is corresponding is pixel original gray value x, ordinate is the number of the pixel that the pixel original gray value is identical with horizontal ordinate in the two field picture, is defined as hist (x).
The statistics with histogram module is each pixel of scan image successively, according to formula (1) statistical picture histogram:
M is the vertical coordinate of image in the formula (1), and n is the horizontal coordinate of image, and h is picture altitude, and w is picture traverse, and x is the gray-scale value of image, and img (m, n) is the view data of the pixel of (m, n) for respective coordinates.
According to histogram distribution relation, calculating upper limit threshold point, middle limit threshold point, lower threshold point, the higher limit between the first Statistical Area, the higher limit between the second Statistical Area.Wherein between the first Statistical Area, between the second Statistical Area according to the higher limit of statistical threshold backstepping interval range, this interval range is along with the distribution situation real-time change of the non-background information of image that collects, because the statistical threshold between Statistical Area is fixed, so according to the upper limit threshold point, middle limit threshold point, the higher limit between lower threshold point and the first Statistical Area, higher limit between the second Statistical Area compares, and can judge roughly whether the non-background information of the current image that collects is many.
The threshold point generation module is seen formula (2) according to the total number VH of pixel in the available gray-scale threshold value G TH statistics available gray-scale interval
Wherein x is the gradation of image value, and hist (x) is the histogram of x for gray-scale value, and G_TH is the available gray-scale threshold value, and its G_TH is according to gradation of image scope capable of regulating, and G_TH is less than or equal to the gradation of image scope;
Described threshold point generation module scans the histogram hist (x) of gray-scale value from 0 to K successively, and every run-down histogram carries out integration to histogram, obtains
Described K is the gradation of image maximal value, and i is gray-scale value corresponding to the current histogram that is scanned, 0≤i≤K.
The threshold point generation module obtains threshold parameter according to the total number VH of the pixel in the available gray-scale interval, lower threshold point scale-up factor LR, middle limit threshold point scale-up factor MR, upper limit threshold point scale-up factor HR, available gray-scale threshold value G_TH according to predefined threshold parameter selection rule.
Described predefined threshold parameter selection rule is:
Work as y
I-1≤ VH * LR and y
iDuring＞VH * LR, the value of i is exactly histogram lower threshold parameter L, if L＞G_TH, then L=G_TH; 0＜LR＜1;
Work as y
I-1≤ VH * MR and y
iDuring＞VH * MR, the value of i is exactly limit threshold parameter M in the histogram, if M＞G_TH, then M=G_TH; 0＜MR＜1 and LR＜MR;
Work as y
I-1≤ VH * HR and y
iDuring＞VH * HR, the value of i is exactly histogram upper limit threshold Parameter H, if H＞G_TH, then H=G_TH; 0＜HR＜1 and MR＜HR;
Work as y
I-1≤ SUM1 and y
iDuring＞SUM1, the i value of this moment is exactly C1, if during i=K and y
iDuring≤SUM1, C1 equals K;
Work as y
I-1≤ SUM2 and y
iDuring＞SUM2, the i value of this moment is exactly C2, if during i=K and y
iDuring≤SUM2, C2 equals K;
SUM1＞SUM2。
The partial error of adjacent two two field pictures opposite sex maximal value is asked for partial error's opposite sex maximal value that module is calculated adjacent two two field pictures, and judges according to partial error's opposite sex maximal value whether the data in the overlay area of two two field pictures exist notable difference.As shown in Figure 3, when sensor surface was all dirty, what obtain during empty the collection was image blurring unclear, chooses in any case the overlay area, and adjacent two two field pictures all are more or less the same, and namely the partial error of adjacent two two field pictures opposite sex maximal value is less than normal.Such as Fig. 5, shown in Figure 6, when the fingerprint image that collects was more clear, the partial error's opposite sex maximal value that calculates adjacent two two field pictures was enough large.
As shown in Figure 7, fix a two field picture motionless, another two field picture that moves up and down obtains the distinct coverage region of two two field pictures.Partial data or the total data chosen in the overlay area are calculated partial error's opposite sex value, the one to one point of choosing the participation computing are calculated partial error's opposite sex value of distinct coverage region with reference to formula (3):
dif(u,v)＝∑(slice1(m,n)-slice2(m+u,n+v))
2 （3）
Slice1 represents the previous frame image in the formula (3), slice2 represents current frame image, m, horizontal coordinate and the vertical coordinate of n presentation video pixel, and m, the scope of n is the scope that participates in the pixel of computing in the selected overlay area, and u, v are horizontal relative displacement and the vertically opposite displacement of the central point of adjacent two two field pictures.
As shown in Figure 8, when the side-play amount result of adjacent two two field pictures asked for module and asks for the side-play amount of adjacent two two field pictures, the horizontal relative displacement of the central point of adjacent two two field pictures was horizontal offset, and vertically opposite displacement is vertical offset.The partial error's opposite sex value that obtains when local otherness minimum value has a plurality of identical and adjacent two two field pictures to cover fully is also minimum, and the relative displacement of central point is side-play amount when preferentially selecting two frames to cover fully, and namely horizontal offset is 0, and vertical offset is 0.
The finger of finger detection module is put and is left basis for estimation with finger and be:
A, when having partially wet finger to be placed on to gather on the sensor, the half-tone information of the non-background of image that collects is many, the image background gray-scale value is higher, the gray-scale value of non-background is on the low side, when the half-tone information of the non-background of image was many, then the main close gray-scale value of gradation of image value distribution was 0 zone, referring to Fig. 4.
By comparing the size of the higher limit between lower threshold point and the first Statistical Area, judge whether the following total number of lower threshold point is abundant.If lower threshold is o'clock more than or equal to the higher limit between the first Statistical Area, the expression gray-scale value arrives the total number of pixel of lower threshold point more than or equal to the statistical threshold between the first Statistical Area 0, represent that namely the following number of lower threshold point is abundant, the image that represents current collection has partially wet finger to put, and expression detects finger and puts.
B, when being placed on when gathering on the sensor without finger, the non-background half-tone information of the image that collects is considerably less.When the half-tone information of the non-background of image is fewer, the gradation of image value main zone near the image background value that distributes then, referring to Fig. 1, Fig. 2.
By comparing the size of the higher limit between upper limit threshold point and the second Statistical Area, judge whether gradation of image information is distributed in more than the upper limit threshold point.If the higher limit between the second Statistical Area is greater than the upper limit threshold point, the expression gray-scale value be 0 to total number of upper limit threshold point less than the statistical threshold between the second Statistical Area, be that the presentation video half-tone information mainly is distributed in more than the upper limit threshold point, the image that represents current collection not finger is put or the situation of the dirty background that sensor gathers when a small amount of residual object is arranged, i.e. expression detects finger and leaves;
C is when the image that gathers is that normal finger or partially dried finger are when being placed on the sensor, such as Fig. 5, shown in Figure 6, the non-background half-tone information of the image that collects is many, and adjacent two two field pictures have obvious lines to change, and the partial error's opposite sex maximal value that calculates adjacent two two field pictures this moment is very large; When there is residue on the whole surface of sensor, as shown in Figure 3, the non-background half-tone information of the image that sky collects is many, and adjacent two two field pictures significantly do not change, whole of collecting is image blurring unclear, and the partial error's opposite sex maximal value that calculates adjacent two two field pictures this moment is very little; When the sensor sheet face portion was dirty, as shown in Figure 2, the half-tone information of the non-background value of image that obtains was fewer.
By the size of the higher limit between relatively middle limit threshold point and the first Statistical Area, whether the number in the judgement below the limit threshold point is abundant.In the middle of limit threshold point during more than or equal to the higher limit between the first Statistical Area, the total number of histogram in the expression below the limit threshold point is more than or equal to the statistical threshold between the first Statistical Area, can get rid of the dirty situation of carrying out empty collection of sensor sheet face portion, the partial error of adjacent two two field pictures of the section of declaring opposite sex maximal value and finger are put the size of threshold value again, when partial error's opposite sex maximal value of adjacent two two field pictures is put threshold value greater than finger, representing that two two field pictures exist significantly changes, and carries out empty situation about gathering when all dirty thereby can get rid of sensor surface.If middle limit threshold point is more than or equal to the higher limit between the first Statistical Area, and the partial error of adjacent two frames opposite sex maximum value is when putting threshold value more than or equal to finger, and then expression detects finger and puts.
D, the non-background information of the image that sky collects after finger leaves is considerably less, and the difference of adjacent two two field pictures is little.
Limit during the size that size by the higher limit between limit threshold point in relatively and the second Statistical Area and partial error's opposite sex maximal value of adjacent two two field pictures and finger leave threshold value is judged below the threshold point the non-background information of image whether insufficient and adjacent two two field pictures whether difference is little.If the higher limit between the second Statistical Area is left threshold value greater than partial error's opposite sex maximal value of middle limit threshold point and adjacent two two field pictures less than finger, the insufficient and adjacent two two field picture differences of the non-background information of image in the expression below the limit threshold point are little, i.e. expression detects finger and leaves;
Pointing mobile basis for estimation is: judge whether present frame detects finger and put, if detecting finger puts, judge again whether one of them is not 0 for the horizontal offset of present frame and previous frame image and vertical offset, if one of them is not 0, it is mobile that expression detects finger; Do not put if detect finger, the expression finger is not mobile.
Adopt the utility model to the finger testing result of the fingerprint image of continuous acquisition as shown in figure 10, the left figure of Figure 10-1, Figure 10-2, Figure 10-3 is the finger Mobile sign, middle graph is the fingerprint image of continuous acquisition, and right figure points to detect to put with finger to leave sign.It is mobile that the little lattice representative of the black of its left figure detects finger, and the little lattice representative of the black of right figure detects finger and puts, and white portion represents to detect finger and leaves.
Those skilled in the art can also carry out various modifications to above content under the condition that does not break away from the definite spirit and scope of the present utility model of claims.Therefore scope of the present utility model is not limited in above explanation, but determined by the scope of claims.
Claims (7)
1. point based on gradation of image information and detect to realize it is characterized in that framework for one kind, comprising:
The statistics with histogram module: according to the gray-scale value of each pixel of scan image, the histogram of real-time statistics present frame fingerprint image is used for generating threshold point;
Threshold point generation module: link to each other with the statistics with histogram module, calculate in real time upper limit threshold point H, middle limit threshold point M, the lower threshold point L of present frame and calculate the higher limit C1 of this segment, the statistical threshold SUM2 between the second Statistical Area according to the statistical threshold SUM1 between the first Statistical Area according to histogram distribution and calculate the higher limit C2 of this segment, thereby obtain the non-background information distribution situation of image;
Partial error's opposite sex maximal value of adjacent two two field pictures is asked for module: be used for partial error's opposite sex value of every kind of overlay area of calculating two two field pictures, and ask for partial error's opposite sex maximal value under all coverage conditions, judge whether adjacent two two field pictures there are differences;
The side-play amount result of adjacent two two field pictures asks for module: ask for partial error's opposite sex value of every kind of overlay area that module obtains according to adjacent two two field picture partial errors opposite sex maximal value, find out the different in nature minimum value of partial error under all coverage conditions; The horizontal relative displacement of the central point of two two field pictures is horizontal offset under the coverage condition of partial error's opposite sex minimum value, and vertically opposite displacement is vertical offset; When two frames cover fully and partial error's opposite sex hour, the relative displacement of the central point when preferentially selecting two frames to cover fully is as side-play amount;
Finger detection module: ask for module with side-play amount result that the partial error of threshold point generation module, adjacent two two field pictures opposite sex maximal value is asked for module, adjacent two two field pictures and link to each other, judge the quantity of the non-background half-tone information of image according to threshold point, get rid of the sensor sheet face portion dirty by the situation of error detection; According to the partial error of adjacent two two field pictures opposite sex maximal value, judge the whether property of there are differences of two two field pictures, get rid of sensor surface all dirty space-time collection by the situation of error detection; According to the side-play amount of adjacent two two field pictures, judge whether finger moves.
2. a kind of the finger based on gradation of image information as claimed in claim 1 detected the realization framework, it is characterized in that, described threshold point generation module is according to the total number VH of pixel in the available gray-scale threshold value G_TH statistics available gray-scale interval, with reference to formula (2)
Wherein x is the gradation of image value, and hist (x) is the histogram of x for gray-scale value, and available gray-scale threshold value G_TH is less than or equal to the gradation of image scope, and G_TH is according to gradation of image scope capable of regulating;
Described threshold point generation module scans the histogram hist (x) of gray-scale value from 0 to K successively, and every run-down histogram carries out integration to histogram, obtains
Described K is the gradation of image maximal value, and i is gray-scale value corresponding to the current histogram that is scanned, 0≤i≤K.
3. a kind of the finger based on gradation of image information as claimed in claim 1 detected the realization framework, it is characterized in that, described threshold point generation module obtains threshold parameter according to the total number VH of the pixel in the available gray-scale interval, lower threshold point scale-up factor LR, middle limit threshold point scale-up factor MR, upper limit threshold point scale-up factor HR, available gray-scale threshold value G_TH according to predefined threshold parameter selection rule.
4. detect the realization framework such as claim 1 or 3 described a kind of fingers based on gradation of image information, it is characterized in that, described predefined threshold parameter selection rule is:
Work as y
I-1≤ VH * LR and y
iDuring＞VH * LR, the value of i is exactly histogram lower threshold parameter L, if L＞G_TH, then L=G_TH; 0＜LR＜1;
Work as y
I-1≤ VH * MR and y
iDuring＞VH * MR, the value of i is exactly limit threshold parameter M in the histogram, if M＞G_TH, then M=G_TH; 0＜MR＜1 and LR＜MR;
Work as y
I-1≤ VH * HR and y
iDuring＞VH * HR, the value of i is exactly histogram upper limit threshold Parameter H, if H＞G_TH, then H=G_TH; 0＜HR＜1 and MR＜HR;
Work as y
I-1≤ SUM1 and y
iDuring＞SUM1, the i value of this moment is exactly C1, if during i=K and y
iDuring≤SUM1, C1 equals K; Work as y
I-1≤ SUM2 and y
iDuring＞SUM2, the i value of this moment is exactly C2, if during i=K and y
iDuring≤SUM2, C2 equals K; SUM1＞SUM2.
5. a kind of the finger based on gradation of image information as claimed in claim 1 detected the realization framework, it is characterized in that, the partial error of described adjacent two two field pictures opposite sex maximal value is asked for partial data or the total data that module chooses the overlay area in and is calculated the different in nature value of partial error, referring to formula (3), and from partial error's opposite sex value, find out partial error's opposite sex maximal value under all coverage conditions;
dif(u,v)＝∑(slice1(m,n)-slice2(m+u,n+v))
2 （3）
Slice1 represents the previous frame image in the formula (3), slice2 represents current frame image, m, horizontal coordinate and the vertical coordinate of n presentation video pixel, and m, the scope of n is the scope that participates in the pixel of computing in the selected overlay area, and u, v are horizontal relative displacement and the vertically opposite displacement of the central point of adjacent two two field pictures.
6. a kind of the finger based on gradation of image information as claimed in claim 1 detect to realize it is characterized in that framework, and the finger of described finger detection module is put and left basis for estimation with finger and be:
A, judge the non-background information amount of image that lower threshold point is following according to the threshold point distribution situation, if lower threshold is o'clock more than or equal to the higher limit between the first Statistical Area, the non-background number of image below the expression lower threshold point is more than or equal to statistical threshold between the first gray area, and expression detects partially wet finger and puts;
B, judge according to the threshold point distribution situation whether the non-background half-tone information of image mainly is distributed in more than the upper limit threshold point, if higher limit is greater than the upper limit threshold point between the second Statistical Area, the number below the expression upper limit threshold point is less than threshold value between the second Statistical Area, and expression detects finger and leaves;
C, whether there is some difference to judge adjacent two two field pictures according to partial error's opposite sex maximal value of the non-background information amount of image below the limit threshold point in the judgement of threshold point distribution situation and adjacent two two field pictures, if middle limit threshold point is put threshold value more than or equal to partial error's opposite sex maximal value of the higher limit between the first Statistical Area and adjacent two two field pictures greater than finger, expression detects finger and puts;
D, judge according to the threshold point distribution situation whether the non-background information of image concentrates on more than the middle limit threshold point, and judge the otherness that adjacent two two field pictures exist according to the partial error of adjacent two two field pictures opposite sex maximal value, if the higher limit between the second Statistical Area is left threshold value greater than partial error's opposite sex maximal value of middle limit threshold point and adjacent two two field pictures less than finger, judge that detecting finger leaves.
7. a kind of the finger based on gradation of image information as claimed in claim 1 detected the realization framework, it is characterized in that, described finger detection module to pointing mobile basis for estimation is: put in the situation that detect finger, judge whether one of them is not 0 for the horizontal offset of present frame and previous frame image and vertical offset, if one of them is not 0, it is mobile that expression detects finger; Do not put if detect finger, it is mobile that expression does not detect finger.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201220477535.4U CN202748806U (en) | 2012-09-19 | 2012-09-19 | Image gray level information-based finger detection realization architecture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201220477535.4U CN202748806U (en) | 2012-09-19 | 2012-09-19 | Image gray level information-based finger detection realization architecture |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN202748806U true CN202748806U (en) | 2013-02-20 |
Family
ID=47708169
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201220477535.4U Expired - Lifetime CN202748806U (en) | 2012-09-19 | 2012-09-19 | Image gray level information-based finger detection realization architecture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN202748806U (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105874469A (en) * | 2014-01-10 | 2016-08-17 | 高通股份有限公司 | Sensor identification |
| CN105912915A (en) * | 2016-05-27 | 2016-08-31 | 广东欧珀移动通信有限公司 | Fingerprint unlocking method and terminal |
| CN107918750A (en) * | 2016-10-08 | 2018-04-17 | 深圳指瑞威科技有限公司 | A kind of adaptive fingerprint image method of adjustment |
| CN108780493A (en) * | 2016-12-14 | 2018-11-09 | 深圳市汇顶科技股份有限公司 | The detection method and device in wet hand region in fingerprint image |
-
2012
- 2012-09-19 CN CN201220477535.4U patent/CN202748806U/en not_active Expired - Lifetime
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105874469A (en) * | 2014-01-10 | 2016-08-17 | 高通股份有限公司 | Sensor identification |
| CN105874469B (en) * | 2014-01-10 | 2017-07-07 | 高通股份有限公司 | Sensor is recognized |
| US10146929B2 (en) | 2014-01-10 | 2018-12-04 | Qualcomm Incorporated | Sensor identification |
| CN105912915A (en) * | 2016-05-27 | 2016-08-31 | 广东欧珀移动通信有限公司 | Fingerprint unlocking method and terminal |
| CN105912915B (en) * | 2016-05-27 | 2017-10-24 | 广东欧珀移动通信有限公司 | A kind of unlocked by fingerprint method and terminal |
| CN107918750A (en) * | 2016-10-08 | 2018-04-17 | 深圳指瑞威科技有限公司 | A kind of adaptive fingerprint image method of adjustment |
| CN108780493A (en) * | 2016-12-14 | 2018-11-09 | 深圳市汇顶科技股份有限公司 | The detection method and device in wet hand region in fingerprint image |
| CN108780493B (en) * | 2016-12-14 | 2022-01-21 | 深圳市汇顶科技股份有限公司 | Method and device for detecting wet hand area in fingerprint image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102103753B (en) | Use method and the terminal of real time camera estimation detect and track Moving Objects | |
| CN202748806U (en) | Image gray level information-based finger detection realization architecture | |
| CN104749184A (en) | Automatic optical detection method and automatic optical detection system | |
| CN108460362A (en) | A kind of system and method for detection human body | |
| CN105163110A (en) | Camera cleanliness detection method and system and shooting terminal | |
| CN107578410B (en) | Automatic identification method for notch image of point switch | |
| CN114549441B (en) | Straw defect detection method based on image processing | |
| CN113674273A (en) | Optical detection method and system based on product defects and readable storage medium | |
| CN105869175A (en) | Image segmentation method and system | |
| CN108710837A (en) | Cigarette smoking recognition methods, device, computer equipment and storage medium | |
| CN114820594A (en) | Method for detecting edge sealing defects of plate based on image, related equipment and storage medium | |
| CN112730251A (en) | Device and method for detecting color defects of screen | |
| CN105991997A (en) | Capacitive image sensor dead pixel real-time positioning method | |
| CN114972173A (en) | Defect detection method, defect detection device and system | |
| CN110740314B (en) | Method and system for correcting defective pixel of color line array camera | |
| CN103679122A (en) | Finger detection framework based on image grey information | |
| CN104749801A (en) | High-precision automatic optical detection method and high-precision automatic optical detection system | |
| CN113283439B (en) | Intelligent counting method, device and system based on image recognition | |
| JP2021052238A (en) | Deposit detection device and deposit detection method | |
| CN105744184B (en) | Bad pixel correction method and the device for using this method | |
| CN103297717A (en) | Self-adaption image processing method and relevant device thereof | |
| JP4918615B2 (en) | Object number detection device and object number detection method | |
| CN103679666B (en) | Improve the system of sensor image quality | |
| CN110349133A (en) | Body surface defect inspection method, device | |
| CN116012322A (en) | Camera dirt detection method, device, equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20170502 Address after: 610041 high tech Zone, Chengdu province Tianfu Avenue North Section of the new century global center, No. 1700 E3-1-1603 Patentee after: Hefei formula Electronic Technology Co.,Ltd. Address before: 610041 Sichuan, Chengdu high tech Tianfu Avenue South extension of the high-tech incubator Park, building A-E-3, No. 1 Patentee before: Chengdu FINCHOS Electrical Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| CX01 | Expiry of patent term |
Granted publication date: 20130220 |
|
| CX01 | Expiry of patent term |