摘要
针对痕迹图像的结构特点及传统图像去噪中所存在的问题,提出了一种基于软阈值的小波图像去噪方法.首先将痕迹图像经过小波分解得到一系列不同程度上的子带图像,再根据噪声和图像信号的不同特性,在不同尺度的子带图像上进行软阈值滤波,最后进行小波重构,得到去噪后的图像.实验结果表明,该算法可较好的保留图像信号的边缘特征,改善图像的视觉效果.
In accordance with characteristic of structure and problem of denosing, the article give a kind of method of wavelet make image denosing based on soft threshold. Mark image be decomposed for a series of son image in degree, the second according to noise and different image characteristic, son image be filtered in degree, finally to make wavelet image and gain denosing image. Experiment result indicates the method that has better results than other algorithms in reserving image edge. Image effect of visual sense be improved.
基金
安徽省科技厅国际合作资助项目(0208800)
安徽省教育厅自然科学基金资助项目(2001kj049).
