一种基于边缘检测的图像去噪优化方法  

An Optimized Method for Image Denoising Based on Edge Detection

作　　者：黄剑玲[1] 郑雪梅[2] 

机构地区：[1]上饶师范学院数学与计算机系,江西上饶334000 [2]上饶师范学院物理系,江西上饶334000

出　　处：《计算机仿真》2009年第11期260-263,共4页Computer Simulation

摘　　要：为了消除或衰减存在于图像上的噪声,同时尽可能地保留图像细节,提出基于边缘检测的图像去噪算法。先通过小波边缘检测法求出有噪图像的边缘图像;再通过小波边缘检测方法确定哪些小波系数是图像的边缘特征,这些小波系数将不受阈值去噪的影响,因此,可以只是根据噪声方差来设置去噪的阈值,对原有噪图像进行小波去噪,得到平滑图像;最后,将边缘图像嵌入平滑图像中,得到去噪后的图像。实验结果表明,与普通的小波阈值去噪方法相比,上述算法不但能在有效去噪的同时保留图像的细节信息,而且能提高去噪后图像的峰值信噪比。In order to eliminate or weaken the image noise , as much as possible to retain the image details at the same time, an image denoising algorithm was presented based on edge detection. At first, this method got the edge image by the wavelet method of edge detection. Then those wavelet coefficients of an image corresponding to image edges were detected by the method of wavelet edge detection. The detected wavelet coefficients would be protected from denoising . therefore the denoising thresholds were set only based on the noise variances, and the smoothed image of the input image were got using the wavelet threshold - denoising method. Finally, the edge image was embedded into the smoothed image, and the final denoising image was gained. The experiment results demonstrate that this algorithm can not only denoise effectively, but also keep the detail information. This method can improve the Signal - to- Noise Ratio, compared with the commonly -used wavelet threshold denoising methods.

关 键 词：边缘检测 阈值去噪 小波变换 边缘图像 平滑图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]