基于FPGA的指纹预处理技术  

Fingerprint image preprocessing based on FPGA

作　　者：谭斌[1] 刘平净[1] 李锋[1] 

机构地区：[1]重庆大学通信工程学院,重庆400030

出　　处：《信息与电子工程》2010年第2期190-195,共6页information and electronic engineering

基　　金：国家自然科学基金资助项目(30570473)

摘　　要：指纹图像预处理是指纹识别系统中的重要部分,通常包括归一化、方向滤波、二值化和细化等环节,算法复杂,计算量大。针对这种特点,设计了一种新的指纹图像预处理方法,运用SOPC技术,将预处理过程中耗时大的部分用FPGA实现,耗时小但处理繁琐的部分用软件实现,由下载到FPGA的软核NiosII处理器负责软件的运行和软、硬件的协调。实验结果表明该方法在确保图像质量提高的前提下大幅度提高了处理速度,保证了自动指纹识别系统的实时性,为指纹识别系统的研究和实现提供了新的设计思路和方法。The fingerprint image pre-processing is a very important part of Automatic Fingerprint Identification System（AFIS）. It includes such processes like normalization，orientation filtering，binarization，thinning，etc. The fingerprint image pre-processing algorithm is complex with plenty of computation. This study designed a new fingerprint image pre-processing method by using System On Programmable Chip（SOPC） technology，realized the most of time-consuming processes by using Field Programmable Gate Array（FPGA），and implemented the part of less time-consuming but handling complexly with the software. The soft-core Nios II processor which had been downloaded to FPGA was responsible for running the software and coordinating the software with hardware．Experimental results showed that this method could improve the quality of images and processing speed greatly，thus it ensured the real-time of AFIS and provided a new design thought for the study of AFIS.

关 键 词：预处理 现场可编程门阵列 指纹方向 指纹频率 GABOR滤波 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]