Design And Realization Of Fingerprint Recognition System Based On FPGA

With the improvement of people’s awareness of safety, more and more products only can be used with personal identity authentication. And Automatic Fingerprint Identification System (AFIS) is one of the most reliable means of recognition, is widely studed, discussed and applied. With the development of integrated circuits, portable electronic equipment is smaller, and has a personal fingerprint identification system. With the less power, smaller size, lower price, faster processing, the traditional fingerprint identification system based on DSP or ARM as the main processor can not meet the requirements.But fingerprint identification system implemented on the hardware can be very prefectly to solve the problems, and meet the requirements needed. This paper designs a fingerprint recognition system in hardware based on FPGA.Firstly, it introduces several traditional fingerprint recognition systems. Then it discuses the related technology and the processing algorithm designed in this paper, which are showed under below:1. Pre-processing:In order to overcome the no-idel segmentation results using the segmentating algorithm based on variance or directional diagram, in this paper an improved algorithm based on variance and its gradient is proposed. Enhancement algorithm is to use the template method based on the direction. And there is an improved method for calculating the midpoint direction. It uses the variance of finger texture structure to calculate the point direction. Optimiazing the deleting and reserving template of the fast OPTA (one-pass thinning algorithm), effectively eliminate the burrs and reduce the numbers of iterations.2. Post-processing:In order to get the eigenvalues, a method of building topology construction of fingerprint feature is designed using an adaptive distance. Based on the several presently known kinds of fingerprints matching algorithm, a new matching algorithm is propounded using the topology construction of fingerprint feature.3. It uses FPS200to obtain the fingerprint image. The original image and temporary data in operating process is stored in SRAM. It uses TFT LCD to display results of pretreatment and LCD1602to indicate the processing state. The eigenvalues of fingerprint are stored in an external SD card.All of the above algorithms and interfaces are programmed with Verilog language. They are designed in the hardware circuit, and simulated test vector based on FPGA and Matlab. Throughing the test results comparing the method of this paper with the others and the thephysical verification of the the system preprocessing hardware in the FPGA development board, it can be seen that the pure hardware implementation of this paper has a higher relative advantage in all aspects of real-time, intergration, low power consumption and low price.It applies the demand for portable electronic equipment development.