论文部分内容阅读
研究了相干斑噪声抑制对合成孔径雷达 (SAR)图像分类的影响。分别采用Kuan自适应滤波和小波变换软门限滤波两种方法进行了相干斑噪声抑制 ;对于SAR图像的分类则采用了图像的灰度以及基于灰度级共生矩阵的 4种纹理特征 ,并利用最大似然分类器进行了监督分类。处理结果表明 ,相干斑噪声的抑制尽管可以提高SAR图像的质量 ,但是由于在相干斑噪声得到抑制的同时 ,地物的固有结构信息也受到损失 ,因此分类精度提高甚微 ,在某些情况下甚至有所下降。针对这种情况 ,提出了一种改进的特征提取方法 ,将基于原图像的灰度级共生矩阵提取的纹理特征与滤波后图像的灰度特征进行组合用于分类。实验结果表明 ,改进的特征提取方法提高了SAR图像的分类精度。
The influence of speckle noise suppression on synthetic aperture radar (SAR) image classification is studied. Two methods of Kuan adaptive filtering and wavelet transform soft-threshold filtering are respectively used to suppress the speckle noise. For the classification of SAR images, four kinds of texture features of the image and the gray level co-occurrence matrix are used, Likelihood classifiers were supervised and classified. The results show that the suppression of speckle noises can improve the quality of SAR images. However, the accuracy of SAR images is not improved due to the suppression of the speckle noise and the loss of intrinsic structural information of objects. In some cases, Even declined. In view of this situation, an improved feature extraction method is proposed, which combines the texture features extracted from the original image-based gray level co-occurrence matrix and the gray features of the filtered image for classification. Experimental results show that the improved feature extraction method improves the classification accuracy of SAR images.