Algorithm Research Of Fingerprint Recognition In Gate Prohibition System

Because of uniqueness and invariability, the fingerprint identification is becoming one of the most popular personal authentication technologies, and many fingerprint products have been offered recently. If we make fingerprint recognition apply to the gate prohibition system, it not only can be used to import and export management, but also contribute to the internal management orderly. Also, it will automatically record the coming and going of the staff, restrict the internal staff access to the region, access time, and refuse uninvited guest politely.On the basis of introducing the knowledge of the fingerprint and the fingerprint identification system's components in this paper, we implement the fingerprint identification algorithm in gate prohibition system including preprocessing, minutiae extraction, how to set up fingerprint minutiae database, and minutiae matching algorithm of the image of fingerprints. In the part of image preprocessing, a segmentation algorithm for fingerprint image based on the information of gray was programmed. According to the directional properties of the fingerprint ridges, a model of directional filter was established to enhance fingerprint image and get rid of the original image noises effectively. Use the dynamic threshold method for fingerprint image binarization and design the algorithm to remove noises according to its features. Use the mathematical morphology for fingerprint image thinning at last. In the part of minutiae extraction, a kind of 8-neighbour extraction algorithm was proposed and distributing regulation of various pseudo minutiae in fingerprint image were analyzed. Special algorithms were designed with respect to various noises to delete pseudo minutiae. In the part of minutiae matching, we use the algorithm based on the similar principle of triangle. According to the coordinates of the character triangle, we can get the rotating and shift parameters, thus the template image and input image can be matched.This topic achieved all of the above algorithms at Visual C++ 6.0 software environment. Through the experiments and the analysis of the experimental results, the algorithms achieve the design requirements basically.