一种基于自适应滤波的指纹图像增强算法  

Fingerprint Image Enhancement Based on Adaptive Filter

作　　者：陈桂友[1] 孙同景[1] 雷印胜[2] 

机构地区：[1]山东大学控制科学与工程学院 [2]天津大学精密仪器与光电子工程学院

出　　处：《电子测量与仪器学报》2006年第6期76-80,共5页Journal of Electronic Measurement and Instrumentation

摘　　要：采集到的指纹图像中往往包含噪声,这些噪声对后续的特征提取和匹配会带来非常不利的影响。为了提高指纹特征提取和匹配的鲁棒性,需对指纹图像进行增强,提高脊线和谷线的对比度,同时滤除图像中的噪声。本文提出了一种基于自适应滤波的指纹图像增强算法,首先改进了LinHong提出的方向场计算方法,然后进行基于方向场的第一次粗滤波;在第一次滤波的基础上计算频率场,最后针对每个分块区域进行自适应增强滤波。实验证明,本文提出的增强算法能有效地对指纹图像进行增强。即使是低质量的指纹图像,也能取得较好的增强效果。There are lots of noise in sampled fingerprint image. They have a great impact on minutia extraction and matching in the automatic fingerprint identification system （AFIS）. In order to improve the robustness of minutiae extraction and matching algorithm, it is essential to preprocess the original fingerprint images to enhance the ridge clarity and the contrast of the ridge and valley, and denoise. A fingerprint image enhancement algorithm based on adaptive filter is presented in this paper. We made some improvements to Lin Hong＇s orientation field estimation algorithm firstly. The first simple filtering phase based on orientation field is performed. The frequency field is calculated on the basis of first filter result. The adaptive filter is carried out to every block at last. Experimental results show that this algorithm can enhance the fingerprint image, even if the quality of the fingerprint image is low.

关 键 词：指纹识别 图像增强 自适应滤波 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]