【摘要】 本文提出一种基于小波变换和频域滤波的指纹图像增强方法。首先提出一种基于小波变换的阈值分割方法,有效地分离出指纹图像的前景区和背景区,对前景区进行小波变换,仅利用小波变换的低频子图估计纹线的方向和频率信息,从而抑制了噪音等高频信息对估计值的影响。针对小波变换后低频子图与高频子图的不同特点和增强目标,在低频子图设计Gabor滤波器,而在各高频子图设计方向滤波器,实现图像的频域滤波。最后进行小波重构。实验结果证明,该算法有效地增强了指纹图像的质量。更多还原

【Abstract】 A fingerprint enhancement algorithm based on wavelet transform and filter in frequency domain is put forward in this pa-per. At first a fingerprint threshold segmentation algorithm based on wavelet transform is brought out which separates the foreground and background of the fingerprint image effectively. Wavelet transform is done in the frontground. Only the low-frequency part is used to determine the ridge orientation and frequency information, so that the effect of the high-frequency information as noise could be restrained. Aiming at the different characteristics and enhancement goals between the low-frequency part and the high-frequency parts, Gabor filter is designed in the low-frequency part while orientation filter is designed in the high-frequency parts. At last reconstitute the image. The experimental result shows that this approach can enhance fingerprint image quality effectively.更多还原