Compound fingers and palms line verifies system
Technical field
The invention belongs to physical characteristics collecting fields, and in particular to a kind of compound fingers and palms line verification system.
Background technology
It is increasingly diversified with the function of integrated circuit, belong to spy's product in the past, or the device that can not be obtained at all,
It all gradually popularizes now.Such as the palmmprint scanner similar to fingerprint scanner, it is exactly an example.Past, in tradition
Consumer product market on it is rare, or the Palm Print Recognition System that all do not listened even is listened, also due to integrated circuit palmmprint scans
The appearance of device is increasingly popularized between the user of access control deeply concerned and identification identification;Palm Print Recognition System
Application field is no longer only limitted to government and Security Officer.These devices are the users for ensuring only to obtain mandate
Just can in a computer system or database into line access, moreover, its volume also narrowed down to can put into it is portable
In computer.
Under the increasingly mature background of web development, the networked instruments of palmmprint scanner are ready to appear.However, due to
Network transmission and distal end fingerprint scanner user instruct without professional person, and the picture quality of acquisition is bad, noise
It is excessive.
Invention content
In order to improve the identification degree of online fingerprint collecting, the present invention provides a kind of compound fingers and palms lines to verify system,
Including:Fingers and palms line long-distance video information acquisition subsystem and fingers and palms line verify subsystem, and the fingers and palms line long-distance video information obtains
Subsystem is taken to obtain the fingerprint and palmmprint of personnel to be identified for the mode based on video frame, the fingers and palms line verification subsystem is used
According to the fingerprint and palmmprint verification personnel identity.
Further, the fingers and palms line long-distance video information acquisition subsystem includes:
First correction processing unit, for carrying out the first gray scale school to collected first frame hand image with initial gray
Positive processing, obtains finger-image, the hand image includes the palm image and finger-image for corresponding to left hand and the right hand respectively;
First intermediate image acquiring unit indicates the corresponding finger of each finger for being obtained using finger lines physiological characteristic
First intermediate image in line region;
Second correction processing unit, for carrying out the second gray scale school to collected second frame hand image with the second gray scale
Positive processing, obtains finger-image, the hand image includes the palm image and finger-image for corresponding to left hand and the right hand respectively;
Second intermediate image acquiring unit indicates the corresponding finger of each finger for being obtained using finger lines physiological characteristic
Second intermediate image in line region;
Palmprint image acquiring unit, for carrying out noise reduction process to the first intermediate image and the second intermediate image, obtain through
Cross the palmprint image of noise reduction.
Further, first correction processing unit includes:
First average gray computing unit, for according to first frame image, calculating the gray scale of left hand and right hand hand image
Average value AGray scale=(AThe left palm+AThe right palm)/2;
First finger determination subelement, in hand image, on the basis of finger shape and length, determining with metacarpus
The image of shortest finger determines the image of time short finger as the corresponding figure of little finger of toe as the corresponding image of thumb in image
Picture determines the image close to thumb as the corresponding image of index finger, determines the image close to little finger of toe as nameless corresponding figure
Picture determines remaining finger-shaped image as the corresponding image of middle finger;
First finger root portion crunode determination subelement, for determining finger root crotch position in hand image
For finger root bifurcation;
First segmentation subelement, for left hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as first point in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as second point;According to each finger root of left hand hand image point
Crunode and first point and second point remove left-hand palm image in left hand hand image, obtain left-hand finger image;
Second segmentation subelement, for right hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as thirdly in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as the 4th point;According to each finger root of right hand hand image point
Crunode and first point and second point remove right hand palm image in right hand hand image, obtain right finger image;
First gray average computation subunit, for calculate separately first point, second point, thirdly, centered on the 4th point,
R is the gray average of the neighborhood territory pixel of radius, and 1 × 4 matrix M is constituted with this 4 gray averages;
First finger gray count subelement, for using the image at tip in left hand hand image as each finger of left hand
The corresponding image of point is calculated using the image at tip in right hand hand image as the corresponding image of each finger tip of the right hand with each
In the corresponding image of a finger tip centered on finger tip position, the gray average for the neighborhood that r is radius, with this 10 gray averages
Constitute 10 × 1 matrix N;
First matrix computation subunit, characteristic value A ' the i.e. feature vector a for calculating the matrix that N × M is obtained;
First establishment of coordinate system subelement, the flat square for using first point as origin, establishing left-hand finger image are sat
The coordinate of mark system, second point is daAnd db, using the 4th point as origin, establish the plane right-angle coordinate of right finger image, third
The coordinate of point is d 'aAnd d 'b;
First intersects correction coefficient computation subunit, for calculating each pixel in left-hand finger image and right finger image
Intersection correction coefficient alpha=A ' × (1-AGray scale×(1-x×ed’a/da)/(1-y×edb/d’b)), obtain the left hand after gray correction
With right finger image, wherein x, y is cross, ordinate value of each pixel in the coordinate system of left hand and the right hand respectively.
Further, the first intermediate image acquiring unit includes:
It hangs down according to each finger extending direction along finger tip to finger and palm junction direction for some finger
The fineness of straight lines searches a most thick lines and obtains the line of demarcation to corresponding hand using this lines as line of demarcation
The region of finger tip, the first intermediate image as the corresponding finger.
Further, second correction processing unit includes:
Bias light gray scale adjusting unit, for the bias light gray scale for acquiring hand image to be adjusted to AGray scale/2;
Second average gray computing unit, for according to the second frame image, calculating the gray scale of left hand and right hand hand image
Average value AGray scale 2=(AThe left palm 2+AThe right palm 2)/2;
Second finger determination subelement, in hand image, on the basis of finger shape and length, determining with metacarpus
The image of shortest finger determines the image of time short finger as the corresponding figure of little finger of toe as the corresponding image of thumb in image
Picture determines the image close to thumb as the corresponding image of index finger, determines the image close to little finger of toe as nameless corresponding figure
Picture determines remaining finger-shaped image as the corresponding image of middle finger;
Second finger root bifurcation determination subelement, for determining finger root crotch position in hand image
For finger root bifurcation;
Third divides subelement, for left hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as first point in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as second point;According to each finger root of left hand hand image point
Crunode and first point and second point remove left-hand palm image in left hand hand image, obtain left-hand finger image;
4th segmentation subelement, for right hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as thirdly in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as the 4th point;According to each finger root of right hand hand image point
Crunode and first point and second point remove right hand palm image in right hand hand image, obtain right finger image;
Second gray average computation subunit, for calculate separately first point, second point, thirdly, centered on the 4th point,
R is the gray average of the neighborhood territory pixel of radius, and 1 × 4 matrix M is constituted with this 4 gray averages;
Second finger gray count subelement, for using the image at tip in left hand hand image as each finger of left hand
The corresponding image of point is calculated using the image at tip in right hand hand image as the corresponding image of each finger tip of the right hand with each
In the corresponding image of a finger tip centered on finger tip position, the gray average for the neighborhood that R is radius, with this 10 gray averages
Constitute 10 × 1 matrix N;
Second matrix computation subunit, characteristic value A " the i.e. feature vector b for calculating the matrix that N × M is obtained;
Second establishment of coordinate system subelement, the flat square for using second point as origin, establishing left-hand finger image are sat
Mark system, first point of coordinate is faAnd fb, thirdly for origin, to establish the plane right-angle coordinate of right finger image, the 4th
The coordinate of point is f 'aAnd f 'b;
Second intersects correction coefficient computation subunit, for calculating each pixel in left-hand finger image and right finger image
Intersection correction coefficient alpha=A " × (1-AGray scale 2× (1-x × lg (f ' a/fa))/(1-y × lg (fb/f ' b))), obtain gray scale school
Left hand after just and right finger image, wherein x, y are cross, ordinate of each pixel in the coordinate system of left hand and the right hand respectively
Value.
Further, the second intermediate image acquiring unit includes:
It hangs down according to each finger extending direction along finger tip to finger and palm junction direction for some finger
The fineness of straight lines searches a most thin lines and obtains the line of demarcation to corresponding hand using this lines as line of demarcation
The region of finger tip, the second intermediate image as the corresponding finger.
Further, the palmprint image acquiring unit includes:
First areal calculation subelement, for calculating left-hand palm area B 1 and right hand palm area B2, hand in first frame
Finger root bifurcation line of the boundary of the palm and finger with reference to each hand;
Second area computation subunit, for calculating left-hand palm area B ' 1 and right hand palm area B ' 2 in the second frame,
Finger root bifurcation line of the boundary of palm and finger with reference to each hand;
Finger areal calculation subelement, for the finger to left hand and the right hand, add up each first middle graph respectively
Image planes are accumulated and the area of the second intermediate image, and then obtain the area of the area and C1 and the second intermediate image of the first intermediate image
And C2;
R-matrix computation subunit obtains matrix E for calculating matrix a × b;
Palm area factor computation subunit, the palm area factor for calculating separately left hand and the right hand:pLeft hand is slapped=ln
(r × (B1/ (2 × B ' 1))), pThe right hand is slapped=ln (R × (B2/ (2 × B ' 2)));
Picture noise filtering subunit, for being filtered to palm image, filtering factor β is:
According to the picture noise filter that filtering factor is β, to the first frame palm image and each finger pair of the second frame
The second intermediate image answered carries out exponent filtering, wherein filtering parameter i.e. filtering factor β.
Further, it is 0.08~0.3 that the r value ranges, which are 0.02~0.1, R,.
Further, the R is 3 times of r.
Further, the initial gray is RGB (255,255,255).
Technical scheme of the present invention has the following advantages:
Using background gray scale when actively changing online fingerprint collecting and creatively with the finger of left hand and the right hand, palm
Gray balance technology and by means of finger and palm acquisition similitude the features such as, due to dither frame when reducing online acquisition
Or network transmission signal-to-noise ratio influences the quality of fingerprint collecting image.After tested, discrimination, which is compared, has online acquisition technology
Improve 40% or more.
Description of the drawings
Fig. 1 shows the composition frame chart of compound fingers and palms line verification system according to the present invention.
Fig. 2 shows the composition frame charts of fingers and palms line long-distance video information acquisition subsystem according to the present invention.
Specific implementation mode
As shown in Figure 1, the compound fingers and palms line verification system of the present invention includes:Fingers and palms line long-distance video acquisition of information subsystem
System and fingers and palms line verify subsystem, and the fingers and palms line long-distance video information acquisition subsystem is obtained for the mode based on video frame
The fingerprint and palmmprint of personnel to be identified, the fingers and palms line verification subsystem are used to verify personnel's body according to the fingerprint and palmmprint
Part.
Preferably, as shown in Fig. 2, the fingers and palms line long-distance video information acquisition subsystem includes:First correction process list
Member obtains finger-image, institute for carrying out the first gradation correction processing to collected first frame hand image with initial gray
It includes the palm image and finger-image for corresponding to left hand and the right hand respectively to state hand image;
First intermediate image acquiring unit indicates the corresponding finger of each finger for being obtained using finger lines physiological characteristic
First intermediate image in line region;
Second correction processing unit, for carrying out the second gray scale school to collected second frame hand image with the second gray scale
Positive processing, obtains finger-image, the hand image includes the palm image and finger-image for corresponding to left hand and the right hand respectively;
Second intermediate image acquiring unit indicates the corresponding finger of each finger for being obtained using finger lines physiological characteristic
Second intermediate image in line region;
Palmprint image acquiring unit, for carrying out noise reduction process to the first intermediate image and the second intermediate image, obtain through
Cross the palmprint image of noise reduction.
Preferably, first correction processing unit includes:
First average gray computing unit, for according to first frame image, calculating the gray scale of left hand and right hand hand image
Average value AGray scale=(AThe left palm+AThe right palm)/2;
First finger determination subelement, in hand image, on the basis of finger shape and length, determining with metacarpus
The image of shortest finger determines the image of time short finger as the corresponding figure of little finger of toe as the corresponding image of thumb in image
Picture determines the image close to thumb as the corresponding image of index finger, determines the image close to little finger of toe as nameless corresponding figure
Picture determines remaining finger-shaped image as the corresponding image of middle finger;
First finger root portion crunode determination subelement, for determining finger root crotch position in hand image
For finger root bifurcation;
First segmentation subelement, for left hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as first point in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as second point;According to each finger root of left hand hand image point
Crunode and first point and second point remove left-hand palm image in left hand hand image, obtain left-hand finger image;
Second segmentation subelement, for right hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as thirdly in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as the 4th point;According to each finger root of right hand hand image point
Crunode and first point and second point remove right hand palm image in right hand hand image, obtain right finger image;
First gray average computation subunit, for calculate separately first point, second point, thirdly, centered on the 4th point,
R is the gray average of the neighborhood territory pixel of radius, and 1 × 4 matrix M is constituted with this 4 gray averages;
First finger gray count subelement, for using the image at tip in left hand hand image as each finger of left hand
The corresponding image of point is calculated using the image at tip in right hand hand image as the corresponding image of each finger tip of the right hand with each
In the corresponding image of a finger tip centered on finger tip position, the gray average for the neighborhood that r is radius, with this 10 gray averages
Constitute 10 × 1 matrix N;
First matrix computation subunit, characteristic value A ' the i.e. feature vector a for calculating the matrix that N × M is obtained;
First establishment of coordinate system subelement, the flat square for using first point as origin, establishing left-hand finger image are sat
The coordinate of mark system, second point is daAnd db, using the 4th point as origin, establish the plane right-angle coordinate of right finger image, third
The coordinate of point is d 'aAnd d 'b;
First intersects correction coefficient computation subunit, for calculating each pixel in left-hand finger image and right finger image
Intersection correction coefficient alpha=A ' × (1-AGray scale×(1-x×ed’a/da)/(1-y×edb/d’b)), obtain the left hand after gray correction
With right finger image, wherein x, y is cross, ordinate value of each pixel in the coordinate system of left hand and the right hand respectively.
Preferably, the first intermediate image acquiring unit includes:
It hangs down according to each finger extending direction along finger tip to finger and palm junction direction for some finger
The fineness of straight lines searches a most thick lines and obtains the line of demarcation to corresponding hand using this lines as line of demarcation
The region of finger tip, the first intermediate image as the corresponding finger.
Preferably, second correction processing unit includes:
Bias light gray scale adjusting unit, for the bias light gray scale for acquiring hand image to be adjusted to AGray scale/2;
Second average gray computing unit, for according to the second frame image, calculating the gray scale of left hand and right hand hand image
Average value AGray scale 2=(AThe left palm 2+AThe right palm 2)/2;
Second finger determination subelement, in hand image, on the basis of finger shape and length, determining with metacarpus
The image of shortest finger determines the image of time short finger as the corresponding figure of little finger of toe as the corresponding image of thumb in image
Picture determines the image close to thumb as the corresponding image of index finger, determines the image close to little finger of toe as nameless corresponding figure
Picture determines remaining finger-shaped image as the corresponding image of middle finger;
Second finger root bifurcation determination subelement, for determining finger root crotch position in hand image
For finger root bifurcation;
Third divides subelement, for left hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as first point in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as second point;According to each finger root of left hand hand image point
Crunode and first point and second point remove left-hand palm image in left hand hand image, obtain left-hand finger image;
4th segmentation subelement, for right hand hand image to be handled as follows:Each finger root bifurcated is clicked through
Row line, and by middle finger and nameless finger root bifurcation and the nameless finger root bifurcation with little finger of toe the two
Metacarpus by middle finger and is eaten close to the intersection point of the profile of little finger of toe as thirdly in the extended line and hand image of the line of crotch
The extended line of the line of the two crotches of the finger root bifurcation and index finger and the finger root bifurcation of thumb of finger with
In hand image metacarpus close to the intersection point of the profile of thumb as the 4th point;According to each finger root of right hand hand image point
Crunode and first point and second point remove right hand palm image in right hand hand image, obtain right finger image;
Second gray average computation subunit, for calculate separately first point, second point, thirdly, centered on the 4th point,
R is the gray average of the neighborhood territory pixel of radius, and 1 × 4 matrix M is constituted with this 4 gray averages;
Second finger gray count subelement, for using the image at tip in left hand hand image as each finger of left hand
The corresponding image of point is calculated using the image at tip in right hand hand image as the corresponding image of each finger tip of the right hand with each
In the corresponding image of a finger tip centered on finger tip position, the gray average for the neighborhood that R is radius, with this 10 gray averages
Constitute 10 × 1 matrix N;
Second matrix computation subunit, characteristic value A " the i.e. feature vector b for calculating the matrix that N × M is obtained;
Second establishment of coordinate system subelement, the flat square for using second point as origin, establishing left-hand finger image are sat
Mark system, first point of coordinate is faAnd fb, thirdly for origin, to establish the plane right-angle coordinate of right finger image, the 4th
The coordinate of point is f 'aAnd f 'b;
Second intersects correction coefficient computation subunit, for calculating each pixel in left-hand finger image and right finger image
Intersection correction coefficient alpha=A " × (1-AGray scale 2× (1-x × lg (f ' a/fa))/(1-y × lg (fb/f ' b))), obtain gray scale school
Left hand after just and right finger image, wherein x, y are cross, ordinate of each pixel in the coordinate system of left hand and the right hand respectively
Value.
Preferably, the second intermediate image acquiring unit includes:
It hangs down according to each finger extending direction along finger tip to finger and palm junction direction for some finger
The fineness of straight lines searches a most thin lines and obtains the line of demarcation to corresponding hand using this lines as line of demarcation
The region of finger tip, the second intermediate image as the corresponding finger.
Preferably, the palmprint image acquiring unit includes:
First areal calculation subelement, for calculating left-hand palm area B 1 and right hand palm area B2, hand in first frame
Finger root bifurcation line of the boundary of the palm and finger with reference to each hand;
Second area computation subunit, for calculating left-hand palm area B ' 1 and right hand palm area B ' 2 in the second frame,
Finger root bifurcation line of the boundary of palm and finger with reference to each hand;
Finger areal calculation subelement, for the finger to left hand and the right hand, add up each first middle graph respectively
Image planes are accumulated and the area of the second intermediate image, and then obtain the area of the area and C1 and the second intermediate image of the first intermediate image
And C2;
R-matrix computation subunit obtains matrix E for calculating matrix a × b;
Palm area factor computation subunit, the palm area factor for calculating separately left hand and the right hand:pLeft hand is slapped=ln
(r × (B1/ (2 × B ' 1))), pThe right hand is slapped=ln (R × (B2/ (2 × B ' 2)));
Picture noise filtering subunit, for being filtered to palm image, filtering factor β is:
According to the picture noise filter that filtering factor is β, to the first frame palm image and each finger pair of the second frame
The second intermediate image answered carries out exponent filtering, wherein filtering parameter i.e. filtering factor β.
Preferably, it is 0.08~0.3 that the r value ranges, which are 0.02~0.1, R,.
Preferably, the R is 3 times of r.
Preferably, the initial gray is RGB (255,255,255).
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement etc., should all be included in the protection scope of the present invention made by within refreshing and principle.