湿地自然边界确定的技术研究——以肃南裕固族自治县为例

The technology of wetland natural boundary determination-Taking Sunan Yugur Autonomous County as an example

本文以肃南裕固族自治县为例,基于高分辨率卫星影像,采用面向对象的最邻近法分类方法准确提取湿地的自然边界,并在提取过程中,利用数字高程模型(DEM)、坡度和地形位置指数(TPI)等地形特征参量,对试验区湿地潜在分布区进行信息提取。结果表明,该方法提取到的湿地自然边界中误差与平均距离均可满足实际生产的需要,提高了分类效率,又最大程度地避免了最邻近法过度依赖样本的缺点。

This paper takes Sunan Yugur Autonomous County as an example,based on high resolution satellite images,adopts object oriented nearest neighbor method to extract the natural bounday of wetlands.The topographic feature parameters such as DEM,slope and TPI have been used in the extraction process to extract effective information about the potential distribution areas of wetlands in the test area.The results show that both the error and the average distance of the wetlands natural boundary extracted by this method can meet the requirements of actual production,improve the classification efficiency and avoid the shortcomings of the nearest neighbor method which relies on samples excessively to the greatest extent.