Fingerprint Identification System Based On DSP

As the main method for figure identification, the technology of fingerprint identification indicates broad application foreground. Taking full advantage of the fingerprint's universal, unique and permanent properties, it has taken place of traditional identification method and was fully used in network, bank, security, treatment and insurance at present. Normally, fingerprint identification mainly includes two aspects on algorithm and application. Along with the development of image processing, many algorithms for fingerprint emerge in endlessly, such as frequency filter, time filter, orientation graph filter; on the other hand, along with the appearance of DSP(Digital Signal Processor) and its development, general DSP chips were more and more used in fingerprint image processing. Combined with these two technology foreground, how to realize rapid and miniaturized identification system has became one of the key problems of research in this domain.According to the background mentioned above, the object researched in this paper is to find out an appropriate algorithm for fingerprint image processing based on DSP technology and develop a fingerprint identification system based on DSP and to realize safe, efficient and convenient personal figure identification. About algorithm design aspect, the theory and basic process of fingerprint identification firstly were analyzed and studied and the new fingerprint identification method--fingerprint image filter based on orientation graph was emphasized secondly in this paper. This algorithm selected eight basic orientations, calculated the gray value varieties of every pixel along these eight orientations, found out the orientation corresponding to minimum variety, respectively constructed point and segment direction graph, chose suitable directional threshold to slit the image, filtered the image using different directional filter templates. The contrastively experimental results show that the image preprocessed with this algorithm has a clear structure and retains the feature of thread and direction. About hardware design aspect, the whole system hardware frame was given firstly, parts including CCD imaging circuit, FPGA time sequence control circuit, DSP nuclear algorithm etc. were described in detailed secondly in this paper. System core is DSP chip made in TI Company. Using its powerful ability on data processing, system can quickly complete kinds of tasks assigned by processing algorithm for fingerprint image and was meet to the request on real time characteristic; with video decoding chip, fingerprint image capture circuit accomplished image decoding, realized interface with high speed DSP component by FIFO, controlled the time sequence circuit depending on FPGA; lastly, in the upper computer, the fingerprint feature database was set up, the final recognition and matching were finished.Proved via experiments, the subject research has certain reference value for the product development and application of embed fingerprint identification system.