基于NSCT和边缘检测的遥感图像去噪

Remote Sensing Image Denoising Based on Edge Detection

噪声对遥感图像的后续处理产生负面影响,为了有效消除遥感图像中的噪声,提出一种基于边缘检测的遥感图像去噪方法。首先对含噪的原始遥感图像进行NSCT分解,得到不同尺度和方向的子带图像,然后利用KPCA对分解后的遥感子带图像进行去噪处理,并采用NSCT对遥感子带图像进行重构,最后采用双边滤波器对去噪后的遥感图像边缘进行平滑处理,消除边缘中的奇异噪声点。实验结果证明,本文方法不仅可以有效去除遥感图像中的噪声,可以保留遥感图像的边缘信息,而且获得了比其它方法更优的遥感图像去噪效果。

Noise has a negative effect on the subsequent processing of remote sensing images,in order to eliminate the noise in the remote sensing images, a method of remoting sensing image denoising based on edge detection is pro- posed. Firstly, the NSCT decomposition of the original remote sensing image with noise is decomposed, and the sub ima- ges of different scales and directions are obtained, and then use KPCA to have depressing noise process with decom- posed sub images, NSCT is used to reconstruct the remote sensing image. Finally, the edge of the remote sensing image is smoothed by bilateral filtering and eliminate the strange noisy point. Experimental results show that the proposed method can effectively remove noise in remote sensing images, and can preserve the edge information, and obtain better results than other methods.