Design Of An Automatic Fingerprint Recognition System Based On PC

The thesis exerts image processing and pattern recognition technique synthetically, study the automatic fingerprint identification theory and technology thoroughly and systemically, realize the pivotal algorithms of fingerprint preprocessing, feature extraction and fingerprint matching.We have identified and explored the following issues:(1)Studying in fingerprint preprocessing. To improve the quality of the recognition, we explore the concept of image evaluating. Aiming at the dark gray level in the center of fingerprint and the low gray level in the suburb of it, we self-adapt the threshold of the block of fingerprint and realize effectively the separation between the ridge and background. Studying in enhancement algorithms of fingerprint, we modify and realize an enhancement algorithm based on Gabor filtering.(2)Studying in the property of Gabor filtering. Because we use it frequently, we discuss the structure and property of Gabor filtering partially.(3)Studying in feature extraction and matching of fingerprint. We discuss the feature extraction technology based on minutiae or texture. At last, we choose a group of filters with 8 Gabor filters to filtering the surround of each reference point. Then we can get a code to indicate fingerprint matching. It is proved to be good.(4) We realize our automatic fingerprint identification system, design the experiment scheme, and test the performance of it.This thesis has research innovation in the aspects of ridge direction extraction, fingerprint image enhancement and fingerprint matching algorithms. And we set up one-one fingerprint identification system. It provides the essential foundation of theoretic and technology for the fingerprint identification system that has the commercial values.