Background technology
Along with infotech develop rapidly, human society are constantly progressive, renewal, higher requirement have been proposed infotech.The demand application that the network information epoch are identified people's identity is more and more, more requires digitizing and the recessivation of identity, how accurately to identify a people's identity, and the protection information security is a critical problem that must solve the information age.
Not being traditional marker or indicating knowledge of biometrics identification technology institute foundation, but a kind of technology that relies on human body biological characteristics to carry out authentication, namely handle by computing machine physiological characteristic or behavioural characteristic collection that human body is intrinsic, carry out the technology that personal identification is identified.At present, some biometric characteristics of differentiating for identity 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, palmmprint identification, auricle identification etc. all belong to the biological characteristic of human body surface.In the prior biological recognition technology, fingerprint and palmmprint be destroyed and forgery easily, and auricle detects at live body and has deficiency.Modes of identity of these identification people only rest on observer's " presentation " of biological characteristic, and safety coefficient is lower.
Second generation biometrics identification technology such as finger vein identification technology are based on the biological characteristic of inside of human body.Finger vena identification has become the main flow research direction of the second generation identity identifying technology of numerous developed countries with its advantage at aspects such as speed, stability, security and crypticities.
Present finger vein recognition system has the following disadvantages:
1, mostly present product is local identification, fails to accomplish networking;
2, travelling speed is slow on computers for the algorithm of finger vena identification;
3, the application of finger vena identification still have many be not developed aspect.
Summary of the invention
The objective of the invention is to fail to accomplish networking, the problem that process operation speed is slow, the application scenario is single in order to solve present finger vein recognition system, the finger vein identification method of a kind of network embedded finger vein recognition system and this system is provided.
Network embedded finger vein recognition system of the present invention comprises N cover collecting device 1 and server 2, and N is the positive integer greater than 0, and described N cover 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 demonstration signal input part of display screen 1-3 is connected with the demonstration signal output part of control circuit 1-4.
The control circuit 1-4 of network embedded finger vein recognition system of the present invention adopts the dm6446 chip to realize.
All embed in the control circuit 1-4 of each collecting device 1 of network embedded finger vein recognition system of the present invention finger vena subgraph storehouse and identification module are arranged, embedding in the server 2 has finger vena picture library and identification module, and the finger vein identification method of network embedded finger vein recognition system of the present invention comprises following content:
The workflow of 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 this original vein image is handled, and obtain gathering image; Execution in step B1;
Step B1, Video processing thread receive the collection image that collecting thread sends, and this collection image is handled the finger venous image after obtaining handling; Execution in step C1;
Finger venous image after step C1, master control thread will be handled and the image in the finger vena word bank in this collecting device 1 mate by matching algorithm; Execution in step D1;
Step D1, master control thread judge whether coupling is successful, if judged result is for being execution in step E1; Otherwise, execution in step F1;
Step e 1, will the match is successful the image corresponding identity information as recognition result, and this recognition result is sent to display screen 1-3 show, also issue simultaneously and send server 2, finish this identification;
Step F 1, the finger venous image after will handling send to server 2, and wait for return message; Execution in step G1;
Step G1, judge whether feedback information is recognition result, if judged result is for being execution in step E1; Otherwise, execution in step H1;
Step H1, transmission recognition failures information show output to display screen 1-3, finish this identification;
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 for being execution in step B2; Otherwise, execution in step C2;
Step B2, storage receive the recognition result of each collecting device 1, and store sequence number and the recognition time information of this recognition result, collecting device 1, finish this identification;
Finger venous image information after step C2, the processing that will receive and the image in the finger vena storehouse compare coupling; Execution in step D2;
Whether step D2, judgement mate successful, if judged result is for being execution in step E2; Otherwise, execution in step F2;
Step e 2, image corresponding identity information that will the match is successful send to control circuit 1-4 in the corresponding collecting device 1 as recognition result, storage simultaneously receives the recognition result of each collecting device 1, and sequence number and the recognition time information of storing this recognition result, collecting device 1, finish this identification;
Step F 2, transmission recognition failures information are finished this identification to the control circuit 1-4 in the corresponding collecting device 1;
Matching algorithm among the above-mentioned steps C1 is realized by following steps:
The unique point of step 1, extraction vein image, described unique point comprises end points and crossbar contact, and 3 * 3 zones in the vein image that extracts through image after handling have 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
TransExpression is from p
1To p
8In the process pixel value be 0 and pixel value be the number of times of 1 checker, described N
TransExpression formula be:
P wherein
9=P
1, work as N
TransMore than or equal to 6 o'clock, then think the point of crossing, work as N
TransMore than or equal to 2 o'clock, then think end points;
Step 2, the unique point of using the contrast of MHD distance algorithm to gather image in image and the vein storehouse, finger vein features is a two-dimentional point set of forming with unique point, finish the coupling of a pair of pattern by calculating two Hausdorff similarities between the point set, for two point sets
With
, the distance definition of MHD is:
N
xRepresent the element number among the 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 that collects by the identification module in the identification module among the control circuit 1-4 and the server 2, with move the finger vena identifying on computers and compare, travelling speed improves 10 times, by server 2 and design server module are set, can support many cover collecting devices to work simultaneously, realize networked finger vena identification.
Embodiment
Embodiment one: present embodiment is described in conjunction with Fig. 1, described, the network embedded finger vein recognition system of present embodiment comprises N collecting device 1 and server 2, N is the positive integer greater than 0, and a described N 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 demonstration signal input part of display screen 1-3 is connected with the demonstration signal output part of control circuit 1-4.
The described network embedded finger vein recognition system of present embodiment is in actual application, and the user can use the process of keyboard 1 control procedure, the time-out of implementation procedure, continues operation and function such as withdraws from.Harvester 1-2 has automatic acquisition function, stretches into harvester 1-2 finishes finger vena in second image acquisition work at finger.Display screen 1-3 is used for the explicit user interactive interface, comprises process selection, finger vena information and result.Control circuit 1-4 goes up operation linux operating system, finger vena identifying and finger vena registration process is downloaded in the (SuSE) Linux OS move.Server 2 LAN (Local Area Network) or Internet link to each other with many covers collecting device, support many cover collecting devices to work simultaneously, handle the task of many cover collecting devices in real time, have realized networked finger vena identification.
Embodiment two: present embodiment is described in conjunction with Fig. 1, the difference of present embodiment and embodiment one described network embedded finger vein recognition system is that the control circuit 1-4 of described network embedded finger vein recognition system adopts the dm6446 chip to realize.
Embodiment three: present embodiment is described in conjunction with Fig. 2 and Fig. 3, present embodiment is the vein identification method of embodiment one described network embedded finger vein recognition system, all embed in the control circuit 1-4 of each collecting device 1 finger vena subgraph storehouse and identification module are arranged, embedding in the server 2 has finger vena picture library and identification module, and described vein identification method comprises following content:
The workflow of 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 this original vein image is handled, and obtain gathering image; Execution in step B1;
Step B1, Video processing thread receive the collection image that collecting thread sends, and this collection image is handled the finger venous image after obtaining handling; Execution in step C1;
Finger venous image after step C1, master control thread will be handled and the image in the finger vena word bank in this collecting device 1 mate by matching algorithm; Execution in step D1;
Step D1, master control thread judge whether coupling is successful, if judged result is for being execution in step E1; Otherwise, execution in step F1;
Step e 1, will the match is successful the image corresponding identity information as recognition result, and this recognition result is sent to display screen 1-3 show, also issue simultaneously and send server 2, finish this identification;
Step F 1, the finger venous image after will handling send to server 2, and wait for return message; Execution in step G1;
Step G1, judge whether feedback information is recognition result, if judged result is for being execution in step E1; Otherwise, execution in step H1;
Step H1, transmission recognition failures information show output to display screen 1-3, finish this identification;
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 for being execution in step B2; Otherwise, execution in step C2;
Step B2, storage receive the recognition result of each collecting device 1, and store sequence number and the recognition time information of this recognition result, collecting device 1, finish identification this time;
Finger venous image information after step C2, the processing that will receive and the image in the finger vena storehouse compare coupling; Execution in step D2;
Whether step D2, judgement mate successful, if judged result is for being execution in step E2; Otherwise, execution in step F2;
Step e 2, image corresponding identity information that will the match is successful send to control circuit 1-4 in the corresponding collecting device 1 as recognition result, storage simultaneously receives the recognition result of each collecting device 1, and sequence number and the recognition time information of storing this recognition result, collecting device 1, finish identification this time;
Step F 2, transmission recognition failures information are to the control circuit 1-4 in the corresponding collecting device 1;
Matching algorithm among the above-mentioned steps C1 is by realizing with step:
The unique point of step 1, extraction vein image, described unique point comprises end points and crossbar contact, and 3 * 3 zones in the vein image that extracts through image after handling have 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
TransExpression is from p
1To p
8In the process pixel value be 0 and pixel value be the number of times of 1 checker, described N
TransExpression formula be:
P wherein
9=P
1, work as N
TransMore than or equal to 6 o'clock, then think the point of crossing, work as N
TransMore than or equal to 2 o'clock, then think end points;
Step 2, the unique point of using the contrast of MHD distance algorithm to gather image in image and the vein storehouse, finger vein features is a two-dimentional point set of forming with unique point, finish the coupling of a pair of pattern by calculating two Hausdorff similarities between the point set, for two point sets
With
, the distance definition of MHD is:
N
xRepresent the element number among the point set X.
The described method of present embodiment can be used needs the finger vein identification technology field, for example: can be applied to technical fields such as gate inhibition, work attendance.Be the example explanation to be applied to the work attendance technical field, in the present embodiment, collecting device at first mates image in the finger venous image that collects and the vein storehouse according to the finger vena subgraph storehouse of its inside, if the match is successful, then draw recognition result, confirm server ip address, found the nested word of socket, this recognition result is sent to server 2, and server 2 keeps this identification record, and the record recognition time, server 2 compares recognition time and Preset Time, whether draw the user late, result such as leave early or do not turn out for work in the database of every day, and is presented at the result outcome record on the display screen 1-3 by control circuit 1-4; If coupling is unsuccessful, finger venous image after handling is sent to server 2, server 2 compares coupling with the image in this image and the finger vena storehouse, if the match is successful, then server 2 keeps this identification record, the collecting device sequence number, and record recognition time, server 2 compares recognition time and Preset Time, draw results the such as whether user is late, leaves early or does not turn out for work, outcome record in the database of every day, and is presented at the result on the display screen 1-3 by control circuit 1-4; If it fails to match, then recognition failures information is presented on the display screen 3 by the control circuit 1-4 in the corresponding collecting device 1.
The finger vein identification method of the described network embedded finger vein recognition system of present embodiment supports that by setting up server 2 and design server identification module many cover collecting devices are worked simultaneously, the while process information.At the server end database of can leafing through at any time, and be provided with function of search, can understand the situation of turning out for work.And real time data can be downloaded to control circuit 1-4, realized the real-time support function of multitask.
Embodiment four: present embodiment is described in conjunction with Fig. 2, the difference of the finger vein identification method of present embodiment and embodiment three described network embedded finger vein recognition systems is that the Video processing thread among the described step B1 to gathering the treatment of picture process is:
Carry out a position cutting to gathering image, take out the position of finger vena in the image;
Image is carried out proportional zoom one time, make the size of image be fit to codec algorithm bag;
Call finger vena codec algorithm bag and handle image, the finger venous image that obtains handling;
Finger venous image is sent to show process and master control process.
The finger vein identification method of the described network embedded finger vein recognition system of present embodiment, for each the frame vein image that collects, can realize the processing time below 30ms, improve applicability and the efficient 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 embodiment four described network embedded finger vein recognition systems is that the process that described codec algorithm bag is handled image may further comprise the steps:
Steps A 3, image is carried out gray scale normalization;
Step B3, image is carried out the texture on the vein direction strengthen;
Step C3, image binaryzation is handled;
Step D3, image is gone hot-tempered point, filtering and refinement.