区域指纹图像增强技术的研究

Research on fingerprint image enhancement in domain

指纹识别技术是指通过计算机,利用人体所固有的指纹生理特征来进行个人身份鉴定的技术。它本质上是模式识别技术的一个分支。指纹图像增强技术可以有效地加强指纹的脊线特征,确保特征提取算法的性能对指纹图像的质量具有足够的鲁棒性,为指纹细节的提取和匹配识别奠定可靠的基础。依据Gabor滤波器能够同时对图像局部结构的方向和空域频率进行解析的特性,介绍了一种基于二维Gabor滤波的指纹图像增强方法,提出了利用Soble算子求解方向梯度进而计算出指纹脊线方向的方法和脊线频率的方法。算法经意大利波格尼亚大学构造的指纹库的一个子集Fingdb得到了验证。

Based on the intrinsic physiological property of fingerprint ,the technique in fingerprint identification is used to judge individual status by way of computer which is a branch of pattern recognition technique. Fingerprint enhancement technique enhances ridge line character effectively, insuring property pick-up algorithm enough robustness to the quality of fingerprint image, establishing reliable base for fingerprint detail pick-up and matching. According as Gabor filter will not only utilize the information of orientation, but also information of interval of fingerprint textures in a local region of fingerprint images, a method of fingerprint enhancement is introduced based on 2-D Gabor filter, a method reckoning orientation of ridge line and frequency of ridge line using Soble arithmetic operators is designed. Results are presented on fingerprint database Fingdb.