The Research And Implementation Of ARM Based Fingerprint Recongnition

As the rapid development of information science and technology, with a higher security identification technology has become the first consideration for many systems. The traditional password and token-based identification techniques exists potential safety hazard such as easily loss and theft, whereas the biological characteristics of authentication provides a more reliable solution. Biometric is biological-based technology using information processing technology which combines biological features organically. Among the many biometric technologies, fingerprint recognition technology is the most widely used, which have many advantages such as universal, permanent, collectable, easily realization. So it has become one of the most effective biometric technologies, and widely used in attendance terminals, criminal identification, information security and other fields because of the lower implementation cost compare to the other biometrics.Fingerprint recognition technology extract fingerprint features in the fingerprint image, and then the characteristics of the two fingerprints are compared by a certain algorithm, finally analysis the match results from two fingerprints to decision. The current fingerprint identification has become an effective means of personal identification, now the urgent requirements of information security makes automated fingerprint identification system be a hotspot, even though the market has many of the shape of the fingerprint identification products, but the accuracy and flexibility of fingerprint recognition cannot meet the requirements of practical application because of many factors such as the noise fingerprint image itself, the skin flexibility, and many research results are not yet open for business interests. The ideal fingerprint identification system is still a very difficult research task at present.This thesis uses 32-bit ARM9 core processor S3C2440A as the master controller, and uses the fingerprint sensor chipset-FPS200 to acquire fingerprint image data, achieved the fingerprint data collection through the serial port on PC. Then several fingerprint recognition algorithms were researched, simulated and compared, which include the filtering, sharpening, image binaryzation, directional diagram, and image thinning etc. in preprocessing. This paper also researched the characteristic points extraction and characteristic match algorithm in detail. Finally one of the algorithms was selected and programmed by integrated design environment C++builder.