Research On Hand Vein Recognition System Based On DSP

With the development of society, now people are paying more and more attention to safety, many occasions need to authenticate and identify the identity of the person, so biometric identification technology is becoming more and more important. As a new biometric technology of human hand vein, when the technology is compared with retinal technology, fingerprint technology and iris technology, it has the advantages of internal features, living body identification, high level of security and non-contract, because the hand vein recognition technology is in the beginning stage, so it is very necessary and has practical significance to study it.At present, the hand vein recognition is usually handled by PC, but the size of such a recognition system is too large, In order to solve this problem, we decided to use DSP in hand vein recognition. In this paper, we choose the TMS320DM642 DSP of TI company, this chip is mainly used for image and video processing, its highest frequency can reach 300 Mhz. The core technology of hand vein recognition products in China is mostly purchased from NEXTERN company of Korea, the application of DSP in biometric devices can effectively reduce the cost and ease of products, so this paper designed a hand vein recognition system based on DSP.In this paper, we use the hand vein image library of Dr. Mohamed Shahin and choose 200 images in experiments, we did a deep research on the hand vein recognition system based on DSP. The main research contents of this paper are:(1) The existing hand vein has a lot of shortcomings, in order to solve the shortcomings of fewer hand vein characteristics and lower recognition rates of hand vein recognition, we proposed a fusion method based on feature point distance of hand vein. The method first extracts seven invariant moments, then extracted the endpoints and the intersections of hand vein, finally we let the geometric features and the local features fused, the accuracy rate can reach 96.75%.(2) This paper introduced the hardware platform of hand vein recognition system, it includes the type of system, system structure and module design. By comparing several DSP, we finally select the DM642 processing chip as the core of the system.(3) This paper designed a software framework of the hand vein recognition system. We also introduced some knowledge of the software development environment, DSP/BIOS operating system etc. The above work has laid the groundwork for the completion of DSP algorithm transplantation.(4) This paper designed a hand vein recognition system based on DSP, when we transplanted the two hand vein recognition algorithm to DSP, the algorithm can successfully determine whether the hand vein is the same people and the accuracy rate can reach 93.5%, through the experiment results we can see that the algorithm is practical and reliable.In this paper, we let the seven moment invariants and the Hausdorff distance weighted fusion, finally the transplantation of hand vein recognition algorithm based on DSP was successfully completed.