基于时间序列MODIS―NDVI冬小麦遥感识别.docVIP

基于时间序列MODIS―NDVI冬小麦遥感识别 　　摘要：利用TM影像更新研究区的土地利用数据，提取冬小麦可能出现的区域作为掩膜限定识别范围，从而可以减少其他植被类型信息的干扰；通过选取冬小麦样点，在时间序列NDVI数据中提取纯冬小麦的时序曲线，根据曲线特征构建时相识别模型；在限定的范围内根据识别模型提取冬小麦，进而将两个尺度数据进行综合处理和面积统计，冬小麦面积为268.65×103 hm2；利用统计年鉴数据和随机抽样两种方法进行精度分析，结果显示面积精度为91.56%，位置精度为87.46%。与实地调查和人工解译相比，大大提供了工作效率，减少了工作量，适用于大面积区域尺度的冬小麦监测。 　　关键词：多时相；NDVI；土地利用类型；冬小麦识别 　　中图分类号：TP79 文献标识码：A 文章编号：0439-8114（2017）08-1560-04 　　DOI：10.14088/j.cnki.issn0439-8114.2017.08.040 　　Winter Wheat Remote Sensing Identification Based on Time Series MOIDS-NDVI 　　LIU Jian-feng1， JIA Yu-qiu2， ZHANG Xi-wang2 　　（1.Yellow River Conservancy Technical Institute， Kaifeng 475004， Henan， China；2.Key Laboratory of Geospatial Technology for Middle and Lower Yellow River Regions， Ministry of Education， Kaifeng 475004， Henan， China） 　　Abstract： In this paper， TM image covering the study area is used to update land use data， from which we can identify where winter wheat may be planted. Then a mask is created， which can reduce interference of other vegetation. Based on the selected samples of winter wheat， NDVI time series of the pure winter wheat pixels are extracted from NDVI products. Then an winter wheat identification model is constructed according to the NDVI curve features. Within the limited range， winter wheat will be identified based on the recognition model， and then the two-scale data are processed in a comprehensive way. Statistical yearbook data and random sampling are used to analyze the accuracy. The results show that the winter wheat acreage is 268.65×103 hm2 in the study area， Acreage accuracy is 91.56% and location accuracy is 87.46%. Compared with field surveys and artificial interpretation， it greatly improves the work efficiency and reduces the workload. Due to the low spatial resolution of MODIS， this method is suitable for crop type identification at regional scale in a large area. 　　Key words： multi-temporal； NDVI； land use type； winter wheat identification 　　?b感技术在实现大面积作物长势监测、遥感估产、病虫害防治等农业生产过程中发挥了很大作用，是促进农业