摘要
本文设计并实现了一套完整的指纹图像自动识别算法,该算法首先对指纹图像进行增强包括归一化,分割指纹图像,区分有效区域和无效区域,计算方向图、用方向滤波的方法对指纹图像进行滤波、接着用动态阈值法二值化经过滤波增强的指纹图像,并对二值化图像进行细化及细化后处理,然后提取细化后图像的特征数据,最后将提取的特征数据进行匹配。实验表明该算法运行时间短,识别效率较高,是一种实用的自动指纹识别算法。
In this paper, a full algorithm is proposed for fingerprint identification： First, original fingerprint images is normalized. Second, segmentation is carried out in order to differentiate between valid area and invalid area. Orientation image is estimated, the fin- gerprint images is filtered using orientation filtering. Next the enhanced image is binaried with dynamic threshold, the fingerprint image is thinned by improved OPTA thinning algorithm. Then the minutiae data is extracted from the thinned image. At last, point matching is used for fingerprint matching. Experimental results show that the algorithm spends less time and has high accuracy.
出处
《微计算机信息》
2012年第6期176-178,168,共4页
Control & Automation
基金
基金申请人:赵磊
项目名称:自动指纹识别算法理论与方法研究
基金颁发部门:湖南省高等学校科学研究项目(10C0172)
衡阳师范学院青年科学基金项目(09A39)
关键词
归一化
方向滤波
细化
特征提取
匹配
normalizing
orientation fdtering
thinning
minutiae extraction
matching
