基于纹理与成像知识的高分辨率SAR图像水体检测

Water detection with high-resolution SAR image based on texture and imaging knowledge

为了精确地提取水体,首先对高分辨率合成孔径雷达(SAR)图像中农村民房的成像机理进行了分析,接着通过图像纹理检测方法和基于知识的建筑物阴影表达去除了建筑物阴影,最后精确地提取出SAR图像中的水体,说明对高分辨率SAR图像进行信息检测时,传统方法和基于成像机理的知识表达方法能够精确地实现水体目标检测,从而可以进一步对水体进行分析。

In order to extract water accurately, we analyze the characteristics and the imaging mechanism of the country buildings in the synthetic aperture rader （SAR） images, and find that the water detection is interfered by the shadow of the buildings. Therefore, the water is accurately extracted from the high-resolution SAR image after getting rid of the building shadows by the texture method and the knowledge-based building representation. This paper indicates that it can more accurately extract information from the high-resolution SAR images with the traditional methods, combining with the knowledge-based methods. After extracting water we can analyze the correlative information farther.