The Pretreatment Algorithm Research And Hardware Design Of A Stand-alone Fingerprint Verification System

As one of the biometric identification technologies, fingerprint identification technology has been applied widely in the society. Along with the development of the embedded system technology, the research of the embedded fingerprint identification technology becomes more and more. In this thesis, the pretreatment algorithm and hardware platform of a stand-alone fingerprint verification system are mainly discussed.Fingerprint identification generally includes fingerprint image pretreatment, fingerprint minutiae extraction, fingerprint minutiae matching. Fingerprint identification has two modes for different applications. One is 1:1 matching, the other is 1:N matching. The 1:1 matching is a process for verification and the 1:N matching is a process for identification. The system acquires the fingerprint of the user on the scene and draws the fingerprint registered before from the fingerprint database according to the user's ID. Then, the system makes a 1:1 matching between the two fingerprints to verify the user's identity.The main contents of this thesis are the pretreatment algorithm and hardware platform of the system. In the aspect of the algorithm research, the fingerprint image acquired by MBF200 is analyzed. In order to remove isolated noise holes and isolated noise blocks, eliminating templates and track following are used for enhancement, other eliminating templates and bifurcation adjustment are used to improve the performance of removing spurious minutiae. In the aspect of the hardware platform design, the requirement and cost of the system is considered. The system can be divided into two parts. One is system control module based on AT89C51, the other is fingerprint processing module based on TMS320C5402. The system control module includes keyboard module, voice playing module, liquid crystal displaying module and information store module. The fingerprint processing module includes fingerprint acquisition module and memory expansion module. The work flows of the two modules and their sub-modules are described and the communication protocol between the two modules is given. As is shown from the result, this algorithm has good effect on eliminating isolated noise holes and isolated noise blocks, also improves the effect of removing spurious minutiae. The research of the stand-alone fingerprint verification system provides good foundation for the network fingerprint verification system.