一种基于纹理特征的卫星遥感图像云探测方法

A Method for Detecting Cloud in Satellite Remote Sensing Image Based on Texture

针对卫星遥感图像中的云探测问题,提出了一种有效的纹理特征分析方法。该方法使用反映图像灰度性质和空间关系的分形和灰度共生矩阵两类纹理,分别来描述云区域和无云的下垫面区域,并从这两个纹理的共5个特征参数构成的多维空间中简化出最小的二维分类空间,使该空间能够完全地区分卫星遥感图像中的云和各种下垫面,利用它设计的线性分类器可以高效地实现云的自动探测功能。大量实际图像测试结果正确率达到98%,证实了该方法的有效性。

A valid texture feature analytical method is proposed to detect cloud in satellite remote sensing image. Using the fractal and gray level co-occurrence texture features to reflect image＇s gray property and space relationship between cloud and the earth surface, from their total five feature parameters this method found a simple two-dimensional classification space which is constructed by the fractal dimension value and energy value of the gray level co occurrence matrix. This space could completely distinguish cloud from various earth surface kinds and the classification implement designed by it realized efficiently cloud automated detection. Numbers of practical images are tested with a detection precise rate of 98% which proved its validity and robustness. This method could be applied for pattern recognition and classification of satellite images under complex background.