指纹图像的自适应预处理研究  

Adaptive pre-processing for fingerprint image

作　　者：薛俊韬[1] 王树成[2] 刘正光[1] 

机构地区：[1]天津大学电气与自动化工程学院,天津300072 [2]天津电信公司,天津300073

出　　处：《计算机工程与设计》2008年第1期157-159,共3页Computer Engineering and Design

摘　　要：自动指纹识别系统的性能在很大程度上依赖于其预处理效果的好坏。不同环境、不同手指采集到的指纹图像,在图像质量方面会有很大的差别,而图像质量的好坏往往决定了识别的准确率。自适应预处理方法首先对指纹图像进行质量评估,对于质量过差的图像直接剔除,不进入后续的处理和识别;对于质量较差的图像,则根据其类型的不同,运用数学形态学和图像处理技术进行自适应增强处理。由于该自适应预处理方法的步骤和结构元素的选择充分考虑到各指纹图像在脊谷线连通性方面的差异,所以对不同的低质量指纹图像,均可得到较好的处理效果。The performance of automatic fingerprint identify system （AFIS） is heavily dependent on the effect ofpre-processing. The fingerprint images collected under different circumstances and fingers have large difference in the quality. The quality of the fingerprint images determines the accurateness of identification generally. An adaptive preprocessing method is proposed. Firstly, the quality check is used as a criterion. Then, the image, whose quality is very low, is rejected directly and is not delivered to the AFIS. The other bad quality image is processed and enhanced by mathematical morphology and image processing technology according to its types. The processing steps and the structure elements depend on the difference of ridges and volleys, so the processing effect for different low quality fingerprint images is good.

关 键 词：低质量指纹图像 图像增强 预处理 数学形态学 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]