摘要
对指纹图像二值化方法作了探讨,采用一种基于方向图的、自适应的指纹图像二值化方法.该方法充分利用了指纹图像本身方向和灰度变化的特点,在保持指纹特征基本不丢失的情况下,可直接从指纹源图像中得到二值化图像,一次性完成一般图像处理中的平滑、增强、二值化的过程.采用梯度法计算指纹的方向图,通过低通滤波器对方向图校正,以方向图为基础对指纹图像进行二值化.实验表明,此算法对质量差的指纹能够取得令人满意的处理效果.
This paper presents an adaptive binarization algorithm for fingerprint images based on orientation map. The method makes full use of orientation and characteristics of grayscale change, gets binarized images from primitive fingerprint images directly, instead of series of processing such as smoothness, enhancement and binarization. Orientation map is firstcomputed by gradients, corrected with low--pass filter, and fingerprint image binarized by orientation map. Experimental results show that this algorithm is effective, especially for fingerprint of poor quality.
出处
《北京师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2009年第3期250-253,共4页
Journal of Beijing Normal University(Natural Science)
基金
北京市教育委员会科技发展计划基金资助项目
关键词
指纹
方向图
梯度法
二值化
fingerprint
orientation map
gradients
binarization
