论文部分内容阅读
为了高精度地提取合成孔径雷达(SAR)图像中的有用信息,提出一种基于灰度共生矩阵的纹理特征辅助SAR图像分类方法,该方法选择的是在合适的窗口尺寸下能将各种地物类型区分开的最佳纹理特征组合.采用增强的Frost滤波法对SAR图像进行斑点噪声抑制,通过比较各典型地物基于灰度共生矩阵的纹理特征统计量,确定参与分类的最佳纹理特征组合、计算灰度共生矩阵的最佳窗口尺寸;采用主成分分析法去除各纹理特征之间的相关性,选择信息量大的2个主成分与图像的灰度共同组成3个波段的图像;最后采用最大似然分类法对该组合图像进行分类.结果表明:该方法提取出的纹理特征辅助SAR图像分类,比无纹理信息参与的SAR图像分类,其精度可提高11.20%.
In order to extract useful information from synthetic aperture radar (SAR) images with high accuracy, a texture feature aided SAR image classification method based on gray level co-occurrence matrix is proposed. The method chooses a method which can classify all kinds of SAR images The optimal texture feature combination is obtained.The enhanced frost filtering method is used to suppress the speckle noise of SAR images and the best texture features involved in the classification are determined by comparing the texture features statistics of each typical feature based on the gray level co-occurrence matrix The optimal window size of the gray level co-occurrence matrix is calculated. Principal component analysis (PCA) is used to remove the correlation between the texture features. The two principal components with large amount of information and the gray level of the image are selected to form three bands of images. Finally, the maximum likelihood classification method is used to classify the combined images.The results show that the proposed method can extract SAR images with texture features, which can increase the accuracy by 11.20% compared with SAR images without texture information.