Fingerprint Recognition Algorithm, And Fpga-based Hardware

With the development of image processing and pattern recognition technology, identification techniques based on biometric characteristics become one of the most rapidly expanding high technology. According to the statistic and prediction of biometric market made by IBG( International Biometric Group) , growth rate of the revenues in this field is about 30% to 50% per year. To 2008, global biometric revenues will up to 4.693 billion dollars. Because of the particular security, stability and facileness, identification technique based on fingerprint characteristics just adapt the request of market. Now, fingerprint identification technology is the most widely used identification technique in the biometric identification field, and also is a hotspot for research and application.SOPC ( System On a Programmable Chip) and embedded system is the most popular conception in the electronic design field. NiosⅡ is a RISC CPU soft core designed by Alters provided with pipelining technology and single instruction flow. It can be embedded in FPGA, integrated with user-defined logic, to build a system on chip based on FPGA. Compared with embedded hard core, soft core is more flexible, and the high-speed characteristic of FPGA device exactly meets the time request of fingerprint identification system.This paper makes an in-depth study to each step's algorithm of fingerprint identification technology, selects feasible algorithms to fit the features of NiosⅡ embedded CPU, and optimizes the algorithm, finally, forms an integrated fingerprint identification algorithm, and presents a hardware design plan about fingerprint identification system based on FPGA.Content of the paper mainly include:1. Improve the pretreatment , aftertreatment and matching algorithm, design a new data structure of fingerprint characteristics for fast matching, and present an fingerprint identification algorithm based on minutiaes. Experimental result shows that the algorithm is high speed, low false accepted rate and high reliable. It will suffice for the request of practicality.2. By using NiosII CPU soft core and peripheral equipment IP cores provided by Alters corporation, build a single-chip embedded system, and present a hardware design plan based on FPGA3. Using NiosII development kit, make a preliminary verification of the hardware design plan, implement the interface between the fingerprint collection chip FPS200 and FPGA, and migrate the algorithmic to NiosII.Experimentations proved that the system design plan presented by this paper is practicable. Automatic fingerprint identification system based on FPGA has particular advantages of speed, power, area and expansibility, and has widely development space.