中国农业气象 ›› 2014, Vol. 35 ›› Issue (03): 344-348.doi: 10.3969/j.issn.1000-6362.2014.03.017

• 论文 • 上一篇    下一篇

基于多源数据和决策树估算夏玉米种植面积

李颖,刘荣花,郑东东   

  1. 1中国气象局河南省农业气象保障与应用技术重点实验室/河南省气象科学研究所,郑州450003;2郑州大学水利与环境学院,郑州450001
  • 收稿日期:2013-11-08 出版日期:2014-06-20 发布日期:2015-02-11
  • 作者简介:李颖(1984-),女,河南郑州人,博士,工程师,主要从事遥感技术及其应用研究。Email:walnutclip@163.com
  • 基金资助:

    2011河南省基础与前沿技术研究计划项目;2014年河南省基础研究计划项目(142300410048); 公益性行业(气象)科研专项(GYHY201006041;GYHY200906022)

Summer Maize Planting Area Estimation Based on Multi Source Data and Decision Tree

LI Ying,LIU Rong hua,ZHENG Dong dong   

  1. 1Henan Key Laboratory of Agro meteorological Support and Applied Technique,CMA/Henan Institute of Meteorological Sciences,Zhengzhou 450003,China;2College of Water Conservancy & Environmental Engineering,Zhengzhou University,Zhengzhou450001
  • Received:2013-11-08 Online:2014-06-20 Published:2015-02-11

摘要: 以低空间分辨率遥感数据为主要信息源提取大范围作物种植信息时,为克服混合像元影响,提高提取精度,提出一种基于多源数据和决策树的夏玉米种植面积估算方法。综合FY-3/MERSI数据的时间序列特征和TM数据的中空间分辨率光谱特征,在对作物物候历进行分析的基础上制定多时相决策树规则,融合多时相MERSI数据和TM数据,分析多源融合后NDVI时序数据的光谱特征确定阈值,提取夏玉米种植信息,根据农业统计数据验证种植面积提取精度为95.1%,根据野外调查数据验证位置精度为81.0%。研究结果可为大范围的作物种植信息准确提取提供方法支持。

关键词: MERSI, 多时相, 阈值, 光谱特征, 归一化植被指数(NDVI)

Abstract: In order to overcome mixed pixels in planting area estimation using low spatial resolution remote sensing data,the method of summer maize planting area estimation based on multi source data and decision tree was proposed.The method used both time series features of FY-3/MERSI and spectral features of TM.The rules of decision tree were established according to local phenological calendar in farmland.Then,MERSI data and TM data were merged and the thresholds of decision tree rules were determined through spectral feature analysis of fused NDVI data to extract summer maize planting informationAccording to agriculture statistics,the accuracy by the proposed approach was 95.1%,and based on the field survey data,the accuracy was 81.0%.The proposed approach can provide reference for crop planting area extraction in large area.

Key words: MERSI, Multi temporal, Threshold value, Spectral features, Normalized difference vegetation index(NDVI)