Research On The Key Technologies Of Fingerprint Collection And Identification Based On OMAP5912

With the continuous progress and development of embedded technology, the very important part in biometric technology-fingerprint identification based on the embedded system is also used more and more widely in all areas of society. During the various types of application process, the requirements of the hardware response time, the efficiency, and the recognition algorithm are getting higher and higher. Thus, high-quality fingerprint capture and recognition technology has become as important topic in the current scientific research.This paper combines the embedded key technologies, biometrics, and image processing technology, study and implement the embedded fingerprint capture and recognition system based on the Linux operating system. System use the dual-core processor of TI’s OMAP5912and scratch-type fingerprint sensor AES2510as the core hardware platform, using embedded Linux as the operating system, improved Gabor filters as the image filtering and preprocessing. Using minutiae-based method of feature points’capture, and based on the type of feature points for fingerprint matching.This paper first analyzes the current situation and development trend of the fingerprint identification system, and design the overall system basis on this ideas and programs. Analysis the key modules of the system functions and working principle, design the hardware circuit based on OMAP5912, a variety of interface hardware circuit and the fingerprint collection module base on AES2510and OMAP5912.Then described in detail how to build the embedded software development platform based on the OMAP5912, including the transplantation of the Linux operating system on the OMAP5912, serial and USB interface-driven design, and the application of the DSP/BIOS.At last, described the fingerprint recognition and image processing workflow and principles, described the idea of the establishment of the fingerprint recognition algorithm. Designed the image filter based on improved Gabor algorithm, and the minutiae extraction algorithm. Algorithm was written in C code embedded CCS environment, and simulation under ADS1.2.Debugging and simulation showed that the system is stable, the acquisition of the fingerprint image is clear, matching algorithm is reliable, the whole system error refused enrollment ratio(FRR) and false acceptance rate(FAR) were controlled within the effective range, robustness, and system design is correct effective and practical.