The Research And Design Of Fingerprint Identification System Based On DSP

Because of its universality,uniqueness and permanency etc,fingerprint have been widely applied in publicsecurity,network,bank,commercial, securities,medic-al,treatment andinsurance etc.In recent years, along with the social’s requirements tofingerprint identification system gradual rising and the rapid developmentof electronic information technology,developing a fast, small, accurate andportability automatic fingerprint identification system (AFIS) not only hasurgent demand of the market but also has a broad application prospect.Based on the analysis of research results of fingerprint identificationtechnology at home and abroad,combining the DSP technology,the paperdo research about fingerprint identification system respectively from twoaspects of hardware and software.Software research is mainly refers to the research of imageprocessing, First of all,in the fingerprint image preprocessing stage, thispaper puts forward an improved image segmentation algorithm.Making use of good real-time matlab software,author makes a simulationexperiment about the traditional variance method based on gray levelcharacteristics and the improved image segmentation algorithm. the resultindicates that the improved fingerprint image segmentation can not onlyimprove the image segmentation effect but also shortened the imageprocessing time. Secondly,in the fingerprint feature extraction phase,thispaper do a detailed study about the post-processing of feature extractionand show the necessity of the post-processing of the fingerprint featureextraction. Finally, in feature matching stage, the paper proposes aimproved method based on the image-calibration for feature pointsmatching and use it to finish the follow-up feature points matching. Thesimulation results show that the improved fingerprint feature pointsmatching method can rise the precision of matching.Choosing the DSP of TI company as the core of the systemprocessor,the paper introduces CPLD logic control module to share tasksfrom the main processor and do research and design for the system of DSPprocessing module, the power modules, storage module, display modules and MBF200acquisition module.From the results of matlab simulation experiment,we can includeconclusion that the improved fingerprint image segmentation method andthe improved fingerprint matching method are more suitable for embeddedfingerprint identification system and helpful for the improvement of theperformance of the system.