摘要
针对现有归一化植被指数(NDVI)阈值方法提取冬小麦种植面积受主观因素影响大,建立的模型不具有普适性等问题,该文以Sentinel-2为数据源,冬小麦种植大县———河南省开封市祥符区(原开封县)为实验区,对基于NDVI时间序列建立的冬小麦识别方法进行改进,使用迭代处理方法改进标准时间序列曲线的确定,提取NDVI时间序列的距离、方向特征以及物候期内最大NDVI值作为识别特征参量,运用数学统计方法确定分类模型阈值,据此构建一种改进的时间序列下冬小麦遥感识别矢量模型。结果表明,该方法对冬小麦遥感识别效果较好,提取的冬小麦种植面积Hellden精度达到92.23%。
Aiming at the problem that the existing normalized difference vegetation index(NDVI)threshold method to extract the influence of winter wheat planting area is influenced by subjective factors and the model does not have the problem such as universality,the vector model of winter wheat remote sensing recognition based on time series feature vector was constructed in this paper,with Sentinel-2used as the data source,and the winter wheat recognition algorithm based on NDVI time series improved for the large winter wheat planting county-Xiangfu district of Kaifeng city,Henan province(formerly Kaifeng county),and the iterative processing method used to improve the determination of the standard time series curve,and the distance and direction characteristics of NDVI time series and the maximum NDVI value during the phenological period extracted as the identification characteristic parameters,and the classification model threshold determined by mathematical statistical method.The results showed that this method had a good effect on the remote sensing recognition of winter wheat,and the Hellden accuracy of the extracted winter wheat planting area was 92.23%.
作者
李蕊
李国清
卢小平
张向军
于海坤
LI Rui;LI Guoqing;LU Xiaoping;ZHANG Xiangjun;YU Haikun(Key Laboratory of Spatio-temporal Information and Ecological Restoration of Mines,MNR,Henan Polytechnic University,Jiaozuo,Henan 454003,China;Henan Remote Sensing and Mapping Institute,Zhengzhou 450003,China)
出处
《测绘科学》
CSCD
北大核心
2022年第4期73-79,110,共8页
Science of Surveying and Mapping
基金
灾害环境下快速应急定位组网技术项目(2016YFC0803103)
河南省自然资源厅2021年度自然资源科研项目。
