Research Of Finger Vein Recognition System Based On Embedded System

With the continuous development of information technology,more and more problems are exposed in the traditional finger vein identification system:the difficult development of bare board,the long period and the low clarity of the original digital vein image collected due to inability to locate accurately by Core components like light source,filter as well as camera.The performance of the recognition algorithm is easily affected by image rotation,translation and noise.Therefore,a digital vein identification system based on embedded system is designed to solve the problems existing in the finger vein recognition system,which has some social benefits.Firstly,the embedded development environment based on Opencv is built,and the NFS server,Samba server and the cross-compilation tool chain are installed.The embedded hardware circuits are designed using core board and extended bottom board.Thus,the problems of difficult development of bare board and long period are solved.Secondly,through the in-depth study of finger vein imaging mechanism,a distance measuring device of Core components is designed,thereby solving the difficulties of low clarity of the original digital vein image owing to the lack of accurate location of Core components.Finally,the Sobel operator is used to extract the ROI region effectively from the original image of the digital vein,and then through curvature detection,threshold segmentation,image refinement,expansion and other techniques,the finger vein skeleton is extracted effectively and the registration template is generated.By calculating improved Zernike moments of the image and combining the classical matching algorithms such as Template matching and Hu moment matching,the effective extraction and high accuracy recognition of finger vein features are realized,which effectively inhibits the influence of image rotation and translation on recognition accuracy.The experimental results show that the system not only achieves high accuracy acquisition,wireless remote transmission and intelligent management,but also has higher recognition accuracy.