基于小波域的图像增强算法

针对图像边缘增强和滤噪的矛盾,提出了符合视觉特性的基于小波域的图像增强算法.将图像的小波变换结果看成一幅图像,则在空域中的阶跃边,在小波域的高频通道中表现为屋脊边,而空域中的屋脊边,在小波域高频通道中表现为成奇对称的屋脊边缘对.在高频通道中,用高斯函数或升余弦函数拟合屋脊边,用正弦函数拟合屋脊边对.通过修改拟合函数的参数来改变其形状,从而使反变换回来的图像的边缘斜率提高,进而达到图像增强的目的.由于在对屋脊边的识别中采用了方向信息测度,使算法能很好地处理边缘增强和滤噪的矛盾.分别对测试图像和实际图像进行处理,并与传统的高通滤波方法进行了对比,实验结果表明,本文提出的方法具有更好的视觉效果.

A wavelet-based image enhancement algorithm

Aimed at the contradiction between edge enhancement and noise filtering,a wavelet-based image enhancement algorithm is presented in this paper. The result of wavelet transform of the image can be seen as another image. The ramp edge in the spatial domain is shown as the roof edge at high frequency band of the wavelet domain,while the roof edge is shown as a pair of odd symmetric roof edges. At high frequency band, the roof edge can be approximated with Gaussian or raised cosine function,and the pair of odd symmetric roof edges can be approximated with sine function. Modifying the parameters of these functions,the slope of edge in the image inversely transformed from the wavelet domain can be increased,therefore the image is enhanced. A method of recognizing the roof edge based on orientation information is also proposed in this study. That algorithm can solve the contradiction between the edge enhancement and noise filtering. According to the experimental results,both the test image and the actual image show that this method outperforms the traditional method,Histogram Equalization.