基于GF-1卫星的县域冬小麦面积提取及年际变化监测

Area Extraction and Interannual Variation Monitoring of Winter Wheat in Counties Based on GF-1 Satellite

利用遥感技术提取冬小麦种植面积及年际变化监测,可为农业经济的宏观决策提供可靠的依据.自2013年我国首颗高分卫星发射以来,高分一号卫星数据以其高分辨率的优势成为了我国农业遥感监测的主要数据源之一.本文以高分一号卫星携带的16m空间分辨率的宽视场(wide field view,WFV)数据为主要数据源,以开封县为例,采用Google Earth(以下简称GE)0.3m遥感影像为样本辅助数据,加上人工目视解译选择样本控制点,采用监督分类中的最大似然法提取冬小麦种植面积,并利用样本点对分类结果进行精度验证和年际变化监测.实验数据表明:采用冬小麦生长周期内的11月份到翌年4月份间的一幅高质量GF-1/WFV影像,利用本文给出的处理流程提取冬小麦面积均可达到农业使用的标准,且1月份为开封县冬小麦提取的最佳时像.同时,采用2013年12月29日、2015年2月25日及2016年2月20日的开封县GF1/WFV数据,面积提取精度均达到95%以上,kappa系数均大于0.96,种植面积提取的数据均略小于河南省统计局公布的数据,但两者呈现出的波动趋势是一致的.结果表明以GF-1/WFV影像为主要数据源,GE影像为辅助数据进行县域尺度上的冬小麦面积提取及年际变化监测的方法是有效可行的.

The remote sensing technology was used to extract the winter wheat planting area and interannual change monitoring,which can provide a reliable basis for the macroeconomic decision-making of agricultural economy.Since the launch of China’s first high-resolution satellite in 2013,the high-resolution satellite data have become one of the major data sources for agricultural remote sensing monitoring in China.This paper took the wide field view(WFV)data of 16 meters spatial resolution carried by the high score satellite as the main data source.Taking Kaifeng County as an example,Google Earth(hereinafter referred to as GE)0.3 mremote sensing image was used.The sample-assisted data,coupled with manual visual interpretation of the selected sample control points,were used to extract the winter wheat planting area using the maximum likelihood method of supervised classification,and the sample points were used to verify the accuracy of the classification results and monitor the interannual variation.The experimental data show that using a high-quality GF-1/WFV image from November to April of the winter wheat growth cycle,the winter wheat area extracted using the processing procedure given in this paper can reach the agricultural use standard,and January is the best time for the extraction of winter wheat in kaifeng country.At the same time,using the GF1/WFV data of Kaifeng County on December 29,2013,February 25,2015 and February20,2016,the area extraction accuracy reaches above 95%,and the kappa coefficient is greater than 0.96.The planting area was extracted.The data are slightly smaller than the data released by the Henan Provincial Bureau of Statistics,but the fluctuations in the two are consistent.The results show that using GF-1/WFV image as the main data source,GE imagery as ancillary data for winter wheat area extraction and monitoring at county scale is effective and feasible.