李峰+谢磊+王昊+秦泉+赵红

摘要：为探讨国产高分一号（GF-1）卫星影像在作物面积提取中的适用性，以冬小麦主产区山东省菏泽市为研究区域，利用GF-1卫星携带的多光谱宽幅相机（WFV）16米遥感影像为主要数据源，以菏泽市土地利用类型和野外地面调查数据作为辅助，采用决策树分类法和监督分类—最大似然分类法相结合的方法，通过分区解译方式，分别提取出菏泽市2014和2015年冬小麦种植面积和分布区域，并利用地面样方数据对分类结果进行精度验证，同时开展年际变化动态监测分析。结果表明：以GF-1/WFV 16米影像为主要数据源，将多源信息引入决策树和监督分类模型，进行种植结构复杂地区冬小麦种植面积遥感估算的方法是可行的。GF-1/WFV 16米影像在作物面积遥感监测业务运行中具有较大的开发应用潜力。2014和2015年菏泽市冬小麦位置提取精度分别达到96.5%和96.7%，面积总量提取精度分别达到96.8%和95.0%；遥感提取的两年冬小麦种植面积均略小于官方提供的统计数据，但两者呈现出的变化趋势一致，即菏泽市两年间的冬小麦种植面积呈减少趋势。

关键词：GF-1/WFV影像；冬小麦；决策树分类；监督分类-最大似然法；菏泽市

中图分类号：S127 文献标识号：A 文章编号：1001-4942（2017）08-0139-06

Abstract In order to explore the applicability of Chinese GF-1 satellite images in crop planting area extraction，the winter wheat planting area and distribution in Heze City， which is one of the main production areas of wheat， was extracted and analyzed based on the GF-1 images. The images obtained by the WFV （wide field view） sensor carried on GF-1 satellite with the spatial resolution of 16 m were taken as the main data source， and the land use type of Heze City and the field survey data were taken as auxiliary ones. By using the method of decision tree classification combined with supervised classification-maximum likelihood method， and the partition interpretation mode， the winter wheat area and distribution of Heze region in 2014 and 2015 were extracted.The classification results were verified using the ground sample data， and the annual variation dynamic were analyzed. The results showed that this method was feasible by introducing the multi-source information into decision tree and supervised classification model to estimate the winter wheat area under complex planting structure based on GF-1 /WFV images.GF-1/WFV image has great application potential in the remote sensing monitoring of crop planting area. The extracting accuracy of winter wheat distribution in Heze in 2014 and 2015 reached 96.5% and 96.7% respectively， and that of total planting area reached 96.8% and 95.0% respectively. The winter wheat planting area extracted by remote sensing was slightly lower than that of official statistics in 2014 and 2015， but the change trend was consistent as decreasing during the two years.

Keywords GF-1/WFV image；Winter wheat；Decision tree classification；Supervised classification- maximum likelihood； Heze City

冬小麥是我国最主要的粮食作物之一，及时准确地获取冬小麦种植面积及空间分布信息，对于准确预测冬小麦产量，加强田间生产管理，优化作物种植结构布局，确保国家粮食安全，具有重要意义[1]。

中国官方发布的农作物种植面积数据主要是通过抽样调查和统计部门逐级上报，此类方法不仅耗时耗力而且缺乏空间分布信息[2，3]。……