魏瑞琪 李林峰 仙巍 邵怀勇 汪盾

摘要：以新疆石河子市2014年的MOD13Q1为数据源，利用遥感技术处理获得NDVI时间序列数据集，运用TIMESAT软件集成的拟合方法[非对称高斯函数（A-G）拟合、双Logistic曲线（D-L）拟合和Savitzky-Golay（S-G）滤波法]对石河子市棉花像元的时间序列数据集进行棉花生长曲线拟合对比，最终选取效果较理想的D-L滤波法对石河子市NDVI时间序列集处理，获取石河子棉花生长曲线，分析棉花生长曲线的特点，提取棉花生长的NDVI阈值，进而提取石河子市的棉花种植区域。结果表明，利用TIMESAT软件和时间序列卫星资料提取棉花种植区域效果好、精度高。MODIS数据分辨率适中，成像面积大，利用时间序列卫星资料获取生长曲线进行棉花信息提取，对大范围的棉花种植监测和农业经济指导具有重要意义。

关键词：遥感；TIMESAT；时间序列；NDVI；生长曲线

中图分类号：P23 文献标识码：A 文章编号：0439-8114（2018）04-0105-08

DOI：10.14088/j.cnki.issn0439-8114.2018.04.028

Extracting Cotton Cultivation Regions of Xinjiang Shihezi Utilizing the TIMESAT and Satellite Time-Series Images

WEI Rui-qi1，LI Lin-feng1，XIAN Wei2，SHAO Huai-yong1，WANG Dun1

（1.Key Lab of Information Technology & Application of Land and Resources，Chengdu University of Technology，Chengdu 610059，China；

2.College of Resources and Environment，Chengdu University of Information Technology，Chengdu 610225，China）

Abstract： Applying remote sensing techniques，this paper chose MOD13Q1 in 2014 covering Shihezi City，Xinjiang Uygur Autonomous Region，and uses TIMESAT as well as satellite time-series data to discuss the abstract of vegetations area based on its growth curve. To select the ideal D-L filtering method on dealing with Shihezi NDVI time-series set，to obtains the growth curve of Shihezi cotton and then to analyzes the characteristics of cotton growth curve，to extract NDVI thresholds in the growth of cotton，further to extract the cotton cultivation regions in Shihezi we downloaded & processed MODIS data to obtain NDVI time-series data， adopted TIMESAT integration method[asymmetric Gaussian function（A-G） fitting， double logistic curve （D-L） fitting and Savitzky-Golay（S-G） filtering] on fitting and comparison between Shihezi cotton pixel time-series data and the cotton growth curve. As the results show，utilizing the TIMESAT and satellite time-series data is the most effective and accurate method to extract cotton cultivation regions. MODIS has a medium data resolution with large imaging area，which obtains the growth curve of cotton to be used for the extraction of cotton information utilizing satellite time-series data，is significant on supervising cotton cultivation on a big scale along with directing agricultural economy.

Key words： remote sensing； TIMESAT； time-series； NDVI； growth curve

中国农作物种植区域数据主要是通过各级统计部门逐级上报进行汇总，人为因素干扰大。建立符合中国国情的农业空间统计调查体系，是中国农业作物监测与统计待解决的问题[1]。及时获取大区域、大尺度农作物空间分布信息也是农情遥感监测的核心问题之一[2]。农作物的生长状况与产量，同样是中国各级管理部门一直非常关注的问题。提取农作物种植区域是遥感估产的基础工作之一，对农作物种植区域及产量动态变化的监测是农业生产领域永恒探索和研究的课题。遥感地理信息系统关键技术——遥感图像分类，是大面积农作物遥感估产的重要技术研究课题，在农作物种植区域估测中占有非常重要的地位。……