基于GF-1号的红枣种植面积提取方法

Extraction Method of Chinese Jujube Planting Area Based on GF-1

新疆南疆地区是我国红枣盛产的主要地区之一,精准掌握红枣种植信息有利于为红枣产业的发展做出更加科学客观决策。为精确提取红枣种植面积,本文以阿克苏地区阿拉尔市为例,基于GF-1号16mWFV传感器卫星数据构建红枣种植面积提取方法。选择归一化植被指数(NDVI)、原始波段、主成分等特征,结合野外实测样点,采用决策树分类方法,提取红枣种植面积。结果表明:该提出方法总体分类精度为94.47%,并且有89.41%的用户精度和91.23%的制图精度。该方法为较大区域红枣种植信息的监督提供参考。

There is one of the major production bases of Chinese jujube in Southern Xinjiang of China. Accurate monitoring of jujube planting area is conducive to making more scientific and objective decisions.The objective of this study is to study the extracting method for jujube planting area based on GF1 satellite data with 16m resolution as an example. The emthod of decision tree classification is employed to extract the planting area for jujube after selecting the normalized vegetation index (NDVI), the original band and the main formation and combining the field sample points. The validated results show that the overall classification, user and mapping accuracy are 94. 47%, 89. 41% and 91. 23%, respectively. The proposed method can provide a reference for monitoring the planting information of jujube at the regional scale.