Research On Fingerprint Identification System Based On DSP

Accuracy and recognition speed of fingerprint recognition algorithm are always needing to be improved. Through a complete comparison and research of the existing fingerprint recognition algorithms, this paper came up with a hardware solution.This paper analyzed comprehensively the image pre-processing algorithm, fingerprint image Minutiae extraction algorithm and fingerprint matching algorithm for the fingerprint identification system, improved the existing algorithm.Besides, the paper presented a new set of fingerprint recognition algorithm. During the finger image pre-processing stage, as long as comparing the thinning image with the 8-connected template, the image will turn into a-single-pixel. During the fingerprint feature extraction stage, By observing the combination of the current point,which is extracted from thinning images, and its eight-connected neighborhood point,we can tell the point type. By this way,the special points also can be exacted from the not so thinned images. To classify the fingerprint image before the fingerprint image matching phase so that the calculation is reduced and the computing speed is increased. The simulation results show that optimization algorithm proposed in this paper is feasible to improve the fingerprint recognition algorithm accuracy and recognition speed. For hardware, the paper presented the system made up of a MBF200 fingerprint sensor and the latest ADI BF533 DSP as a core processor for improving system performance.Throughout the experiment, the matching rate was higher and the fingerprint recognition speed was better. In general, the fingerprint recognition system was designed more reasonably and met the design objectives basically.