基于纹理特征的水系信息提取

Extracting the Information of Water System Based on Texture Features

以往的遥感水系信息提取多是基于图像的波谱信息,而水系的构成形态更多展现了空间的变化,所以反映空间变化情况的纹理信息在水系提取方面更有优势。基于图像纹理特征提取水系的方法研究,通过统计图像的灰度共生矩阵,使用对比度和方差统计量对水系信息进行了提取研究,结果证明纹理特征对水系的准确提取非常有效。

Extracting water system information from remote sensing image always with the spectrum character,and the component modality of water system represents the spatial characteristic further,so the texture features is more useful.The authors study the means of extracting water system based on texture features,by compute the Gary Level Co-occurrence Matrix,use the contrast and the variance features extracting the water system information,the result prove it is very effective.