Design And Implementation Of Fingerprint Recognition System Based On Tms320f2812

Fingerprint identification system is one of the most cost-effective biometric systems in the world, fingerprint identification technology and embedded systems can be combined, in this way, the application of fingerprint identification technology can be more widely applicable to various situations without considering space and environmental constraints. As the continuous development of economy and society, identity authentication is needed in more and more occasions, there will be more consumption in fingerprint identification products in our country, therefore, it is significance to develop a cost-effective embedded fingerprint identification system with high identification speed.In this thesis, what we do is summarized as below:First, according to project design requirement, we develop the embedded hardware module of automated fingerprint identification system based on TMS320F2812DSP; secondly, we collect pixels of fingerprint with the fingerprint sensor FPC1011F and process pixels using MATLAB to show the collected fingerprint images; again, we will research and optimize the fingerprint feature extraction and feature point matching algorithms, and verify whether the algorithms are correct in the PC, using VC++. For the special hardware architecture of DSP chip, we will transplant algorithms which have been verified on the PC, build the database of fingerprint features, and realize fingerprint registration, fingerprint matching, uploading and downloading database of fingerprint features by serial port; at last, because codes are executed faster in the internal SRAM, when fingerprint identification module will be on power, parts of the core codes will be loaded from the Flash into the SRAM, in this way, we can greatly enhance the speed of fingerprint feature extraction and matching.Experimental results show that the various parts of the system work properly, the system identification time<1 second, error refused to read rate (FRR)<5%, false acceptance rate (FAR)<1/106, the design meets application needs basically, has reached the expected requirements, we have done the basic research work for the further development of the fingerprint identification system.