高分辨率影像农田信息提取方法

Method for Extracting Farmland Information from High Resolution Satellite Image

选取河南省安阳市某地冬小麦长势良好的Quick Bird影像,首先应用数学形态学的开运算对高分辨率遥感影像进行分割,将种植冬小麦的农田与反射率较高的空地、道路、农村居民地进行区分,人工选择并提取需要分析的农田区域,计算其边界并进行细化,最终提取农田边界信息。整体处理流程基于VC++6.0编程实现。

The Quick Bird images of a certain area in Anyang, Henan province were taken as the testing data in this paper. First of all, high resolution satellite images were segmented by open operation of the mathematical morphology in order to separate fields from open space, roads and rural residential, most of whom have high reflectivity. Secondly, interesting fields were selected manually using the method of seed region grow. At last, the boundary of the field was extracted by boundary tracing and thinned by the mathematical morphology thinning operation. The whole process was actualized by VC programming.