Background technology
Along with the constantly progress of infotech develop rapidly, human society, renewal, higher requirement are proposed to infotech.The application of demand that the identity of Network Information epoch to people identifies is more and more, more requires digitizing and the recessivation of identity, how the identity of a precise Identification people, and protection information safety is a critical problem that must solve the information age.
Biometrics identification technology institute foundation be not traditional marker or indicate knowledge, but rely on human body biological characteristics to carry out a kind of technology of authentication, namely by computing machine, physiological characteristic intrinsic for human body or behavioural characteristic collection are processed, carry out the technology of personal identification qualification.At present, some biometric characteristics for identity verify mainly contain vocal print, fingerprint, face line, iris, person's handwriting, gait, infrared temperature spectrogram etc.
First generation biological identification technology such as fingerprint recognition, personal recognition, Ear recognition etc. all belong to the biological characteristic of human body surface.In existing biological identification technology, fingerprint and palmmprint be destroyed and forgery easily, and auricle is at In vivo detection Shortcomings.These identify that the mode of the identity of people only rests in " presentation " of the biological characteristic of observer, and safety coefficient is lower.
Second generation biometrics identification technology as finger vein identification technology, based on the biological characteristic of inside of human body.Finger vena identification, with its advantage in speed, stability, security and crypticity etc., has become the mainstream research direction of the second generation identity identifying technology of numerous developed country.
Current finger vein recognition system has the following disadvantages:
1, mostly current product is local and identifies, fails to accomplish networking;
2, travelling speed is slow on computers for the algorithm of finger vena identification;
3, still have in the application of finger vena identification many be not developed in.
Summary of the invention
The object of the invention is to fail to accomplish networking, the problem that process operation speed is slow, application scenario is single to solve current finger vein recognition system, a kind of finger vein identification method of network embedded finger vein recognition system is provided.
Network embedded finger vein recognition system of the present invention comprise N overlap collecting device 1 and server 2, N be greater than 0 positive integer, described N is overlapped collecting device 1 and is connected with server 2 by LAN (Local Area Network) or Internet;
Collecting device 1 is made up of keyboard 1-1, harvester 1-2, display screen 1-3 and control circuit 1-4, the USB port of described keyboard 1-1 is connected with the USB port of control circuit 1-4, the image signal output end of harvester 1-2 is connected with the picture signal input end of control circuit 1-4, and the display input end of display screen 1-3 is connected with the display output terminal of control circuit 1-4.
The control circuit 1-4 of network embedded finger vein recognition system of the present invention adopts dm6446 chip to realize.
Finger vena subgraph storehouse and identification module is all embedded with in the control circuit 1-4 of each collecting device 1 of network embedded finger vein recognition system of the present invention, be embedded with finger vena picture library and identification module in server 2, the finger vein identification method of network embedded finger vein recognition system of the present invention comprises following content:
The workflow of the control circuit 1-4 in each collecting device 1 is:
Steps A 1, collecting thread receive the finger vena original image that harvester 1-2 sends, and process this original vein image, obtain gathering image; Perform step B1;
Step B1, Video processing thread receive the collection image that collecting thread sends, and process this collection image, obtain the finger venous image after processing; Perform step C1;
Finger venous image after process is mated by matching algorithm with the image in the finger vena word bank in this collecting device 1 by step C1, master control thread; Perform step D1;
Step D1, master control thread judge that whether coupling is successful, if judged result is yes, perform step e 1; Otherwise, perform step F 1;
Step e 1, using identity information corresponding for the image that the match is successful as recognition result, and sent to by this recognition result display screen 1-3 to show, also issue simultaneously and send server 2, complete this and identify;
Step F 1, the finger venous image after process is sent to server 2, and wait for return message; Perform step G1;
Step G1, judge whether feedback information is recognition result, if judged result is yes, perform step e 1; Otherwise, perform step H1;
Step H1, transmission recognition failures information, to display screen 1-3 display translation, complete this and identify;
The workflow of server 2 is:
Steps A 2, judge whether the information that collecting device 1 sends is recognition result, if judged result is yes, perform step B2; Otherwise, perform step C2;
Step B2, store and receive the recognition result of each collecting device 1, and store this recognition result, the sequence number of collecting device 1 and recognition time information, complete this and identify;
Step C2, the finger venous image information after the process received and the image in finger vena storehouse are carried out contrast mate; Perform step D2;
Step D2, judge coupling whether success, if judged result is yes, perform step e 2; Otherwise, perform step F 2;
Step e 2, control circuit 1-4 identity information corresponding for the image that the match is successful sent to as recognition result in corresponding collecting device 1, store the recognition result receiving each collecting device 1 simultaneously, and store this recognition result, the sequence number of collecting device 1 and recognition time information, complete this and identify;
Step F 2, transmission recognition failures information, to the control circuit 1-4 in corresponding collecting device 1, complete this and identify;
Matching algorithm in above-mentioned steps C1 is realized by following steps:
The unique point of step one, extraction vein image, described unique point comprises end points and crossbar contact, for 3 × 3 region of in the vein image after image zooming-out process, has 9 pixels, and described 9 pixels are respectively p
0, p
1, p
2, p
3, p
4, p
5, p
6, p
7and p
8, the pixel value of described 9 pixels is respectively P
0, P
1, P
2, P
3, P
4, P
5, P
6, P
7and P
8if certain pixel has vein image, then the pixel value of this pixel is 1, otherwise is 0, if P
0=1, N
transrepresent from p
1to p
8in process pixel value be 0 and pixel value be the number of times of 1 checker, described N
transexpression formula be:
Wherein P
9=P
1, work as N
transwhen being more than or equal to 6, then think point of crossing, work as N
transwhen being more than or equal to 2, then think end points;
Step 2, the contrast of use MHD distance algorithm gather the unique point of image in image and vein storehouse, finger vein features is a two-dimentional point set with unique point composition, the coupling of a pair pattern has been carried out, for two point sets by the Hausdorff similarity calculated between two point sets
with
the distance definition of MHD is:
N
xrepresent the element number in point set X.
The finger vein identification method of network embedded finger vein recognition system of the present invention is identified the finger venous image collected by the identification module in the identification module in control circuit 1-4 and server 2, with run compared with finger vena identifying on computers, travelling speed improves 10 times, by arranging server 2 and design server module, can support that many cover collecting devices work simultaneously, achieve the identification of networking finger vena.
Embodiment
Embodiment one: composition graphs 1 illustrates present embodiment, described in present embodiment, network embedded finger vein recognition system comprises N number of collecting device 1 and server 2, N be greater than 0 positive integer, described N number of collecting device 1 is connected with server 2 by LAN (Local Area Network) or Internet;
Collecting device 1 is made up of keyboard 1-1, harvester 1-2, display screen 1-3 and control circuit 1-4, the USB port of described keyboard 1-1 is connected with the USB port of control circuit 1-4, the image signal output end of harvester 1-2 is connected with the picture signal input end of control circuit 1-4, and the display input end of display screen 1-3 is connected with the display output terminal of control circuit 1-4.
Network embedded finger vein recognition system described in present embodiment is in actual application, and user can use the process of keyboard 1 control procedure, the time-out of implementation procedure, continues functions such as running and exit.Harvester 1-2 has automatic acquisition function, stretches at finger the image acquisition work completing finger vena in harvester 1-2 second.Display screen 1-3, for showing User Interface, comprises process choosing, finger vena information and result.Control circuit 1-4 runs linux operating system, finger vena identifying and finger vena registration process is downloaded in (SuSE) Linux OS and runs.Server 2 LAN (Local Area Network) or Internet with how overlap collecting device and is connected, support that many cover collecting devices work simultaneously, the task of collecting device is overlapped in process in real time more, achieves the identification of networking finger vena.
Embodiment two: composition graphs 1 illustrates present embodiment, the difference of present embodiment and the network embedded finger vein recognition system described in embodiment one is, the control circuit 1-4 of described network embedded finger vein recognition system adopts dm6446 chip to realize.
Embodiment three: composition graphs 2 and Fig. 3 illustrate present embodiment, present embodiment is the vein identification method of the network embedded finger vein recognition system described in embodiment one, finger vena subgraph storehouse and identification module is all embedded with in the control circuit 1-4 of each collecting device 1, be embedded with finger vena picture library and identification module in server 2, described vein identification method comprises following content:
The workflow of the control circuit 1-4 in each collecting device 1 is:
Steps A 1, collecting thread receive the finger vena original image that harvester 1-2 sends, and process this original vein image, obtain gathering image; Perform step B1;
Step B1, Video processing thread receive the collection image that collecting thread sends, and process this collection image, obtain the finger venous image after processing; Perform step C1;
Finger venous image after process is mated by matching algorithm with the image in the finger vena word bank in this collecting device 1 by step C1, master control thread; Perform step D1;
Step D1, master control thread judge that whether coupling is successful, if judged result is yes, perform step e 1; Otherwise, perform step F 1;
Step e 1, using identity information corresponding for the image that the match is successful as recognition result, and sent to by this recognition result display screen 1-3 to show, also issue simultaneously and send server 2, complete this and identify;
Step F 1, the finger venous image after process is sent to server 2, and wait for return message; Perform step G1;
Step G1, judge whether feedback information is recognition result, if judged result is yes, perform step e 1; Otherwise, perform step H1;
Step H1, transmission recognition failures information, to display screen 1-3 display translation, complete this and identify;
The workflow of server 2 is:
Steps A 2, judge whether the information that collecting device 1 sends is recognition result, if judged result is yes, perform step B2; Otherwise, perform step C2;
Step B2, store and receive the recognition result of each collecting device 1, and store this recognition result, the sequence number of collecting device 1 and recognition time information, complete and this time identify;
Step C2, the finger venous image information after the process received and the image in finger vena storehouse are carried out contrast mate; Perform step D2;
Step D2, judge coupling whether success, if judged result is yes, perform step e 2; Otherwise, perform step F 2;
Step e 2, control circuit 1-4 identity information corresponding for the image that the match is successful sent to as recognition result in corresponding collecting device 1, store the recognition result receiving each collecting device 1 simultaneously, and store this recognition result, the sequence number of collecting device 1 and recognition time information, complete and this time identify;
Step F 2, transmission recognition failures information are to the control circuit 1-4 in corresponding collecting device 1;
Matching algorithm in above-mentioned steps C1 is by realizing with step:
The unique point of step one, extraction vein image, described unique point comprises end points and crossbar contact, for 3 × 3 region of in the vein image after image zooming-out process, has 9 pixels, and described 9 pixels are respectively p
0, p
1, p
2, p
3, p
4, p
5, p
6, p
7and p
8, the pixel value of described 9 pixels is respectively P
0, P
1, P
2, P
3, P
4, P
5, P
6, P
7and P
8if certain pixel has vein image, then the pixel value of this pixel is 1, otherwise is 0, if P
0=1, N
transrepresent from p
1to p
8in process pixel value be 0 and pixel value be the number of times of 1 checker, described N
transexpression formula be:
Wherein P
9=P
1, work as N
transwhen being more than or equal to 6, then think point of crossing, work as N
transwhen being more than or equal to 2, then think end points;
Step 2, the contrast of use MHD distance algorithm gather the unique point of image in image and vein storehouse, finger vein features is a two-dimentional point set with unique point composition, the coupling of a pair pattern has been carried out, for two point sets by the Hausdorff similarity calculated between two point sets
with
the distance definition of MHD is:
N
xrepresent the element number in point set X.
Method described in present embodiment can be applied needs finger vein identification technology field, such as: can be applied to the technical field such as gate inhibition, work attendance.Illustrate to be applied to work attendance technical field, in present embodiment, first collecting device mates with image in vein storehouse the finger venous image collected according to the finger vena subgraph storehouse of its inside, if the match is successful, then draw recognition result, confirmed service device IP address, found the nested word of socket, this recognition result is sent to server 2, server 2 retains this identification record, and record recognition time, recognition time and Preset Time contrast by server 2, show whether user is late, the result such as to leave early or do not turn out for work, by outcome record in the database of every day, and result is presented on display screen 1-3 by control circuit 1-4, if mate unsuccessful, finger venous image after process is sent to server 2, this image and the image in finger vena storehouse are carried out contrast and mate by server 2, if the match is successful, then server 2 retains this identification record, collecting device sequence number, and record recognition time, recognition time and Preset Time contrast by server 2, draw the results such as whether user is late, leaves early or does not turn out for work, by outcome record in the database of every day, and result is presented on display screen 1-3 by control circuit 1-4, if it fails to match, then recognition failures information is presented on display screen 3 by the control circuit 1-4 in corresponding collecting device 1.
The finger vein identification method of the network embedded finger vein recognition system described in present embodiment by erection server 2 and design server identification module, supports that many cover collecting devices work simultaneously, process information simultaneously.Can to leaf through at any time database at server end, and be provided with function of search, can attendance be understood.And real time data can be downloaded to control circuit 1-4, achieve multitask real-time support function.
Embodiment four: composition graphs 2 illustrates present embodiment, the difference of the finger vein identification method of present embodiment and the network embedded finger vein recognition system described in embodiment three is, the Video processing thread in described step B1 to the processing procedure gathering image is:
A position cutting is carried out to collection image, takes out the position of finger vena in image;
A proportional zoom is carried out to image, makes the size of image be applicable to codec algorithm bag;
Call finger vena codec algorithm bag process image, obtain the finger venous image processed;
Finger venous image is sent to show process and host process.
The finger vein identification method of the network embedded finger vein recognition system described in present embodiment, for each the frame vein image collected, can realize the processing time at below 30ms, improve applicability and the efficiency of finger vena identification, recognition success rate reaches more than 99%.
Embodiment five: the difference of the finger vein identification method of present embodiment and the network embedded finger vein recognition system described in embodiment four is, the process of described codec algorithm bag process image comprises the following steps:
Steps A 3, gray scale normalization is carried out to image;
Step B3, texture on a vein direction is carried out to image strengthen;
Step C3, to image binaryzation process;
Step D3, hot-tempered point, filtering and refinement are gone to image.