主办单位:中国气象局沈阳大气环境研究所
国际刊号:ISSN 1673-503X
国内刊号:CN 21-1531/P

气象与环境学报 ›› 2024, Vol. 40 ›› Issue (6): 89-97.doi: 10.3969/j.issn.1673-503X.2024.06.011

• 论文 • 上一篇    下一篇

基于多时相MODIS数据的黑龙江省大豆种植区识别与面积估算

宋瑞1,2(),赵恒谦2,于文颖1,*(),杨屹峰2,李子涵2   

  1. 1. 中国气象局沈阳大气环境研究所,辽宁沈阳 110166
    2. 中国矿业大学 (北京) 地球科学与测绘工程学院,北京 100083
  • 收稿日期:2023-05-27 出版日期:2024-12-28 发布日期:2025-01-21
  • 通讯作者: 于文颖 E-mail:song13141011@163.com;yuwenying@iaesy.cn
  • 作者简介:宋瑞,女,1999年生,在读硕士研究生,主要从事农业遥感大数据分析方面的研究,E-mail: song13141011@163.com
  • 基金资助:
    国家自然科学基金(41971401);教育部产学合作协同育人项目(202102245005);中国气象局沈阳大气环境研究所联合开放基金(2021SYIAEKFMS41);中国矿业大学(北京)越崎青年学者项目(2020QN07)

Identification and area estimation of soybean planting areas in Heilongjiang province based on multi-temporal MODIS data

Rui SONG1,2(),Hengqian ZHAO2,Wenying YU1,*(),Yifeng YANG2,Zihan LI2   

  1. 1. Institute of Atmospheric Environment, China Meteorological Administration, Shenyang 110166, China
    2. School of Earth Science and Surveying and Mapping Engineering, China University of Mining and Technology, Beijing 100083, China
  • Received:2023-05-27 Online:2024-12-28 Published:2025-01-21
  • Contact: Wenying YU E-mail:song13141011@163.com;yuwenying@iaesy.cn

摘要:

选用2017—2021年黑龙江省多时相MODIS影像数据,基于Google Earth Engine(GEE) 地理空间分析云平台,对比分析各类地物光谱反射率以及归一化植被指数(NDVI)、归一化水体指数(NDWI)等指数差异特征,建立分类决策树,识别黑龙江省大豆种植区、估算面积,并与其他分类方法进行比较。结果表明:基于决策树分类方法识别的2018年黑龙江省大豆种植区面积估算总精度为97.09%,Kappa系数为0.77,分类精度高于随机森林和支持向量机法。通过调整优化决策树模型,进行无样本年份大豆种植区识别和面积估算,得到2017—2021年黑龙江省大豆分布变化,总精度为90%以上,Kappa系数大于0.60,面积估算结果精度达95%以上。

关键词: 植被指数, 水体指数, 光谱分析, 决策树分类

Abstract:

Multi-temporal MODIS image data of Heilongjiang province from 2017 to 2021 were selected, and based on Google Earth Engine (GEE) geospatial analysis cloud platform, the spectral reflectance of various types of features as well as the difference characteristics of indices such as Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI) were compared and analyzed.A classification decision tree was established to identify soybean planting areas in Heilongjiang province, estimate the area, and compare with other classification methods.The results show that the total accuracy of estimating the area of soybean planting areas in Heilongjiang province in 2018 identified based on the decision tree classification method is 97.09%, and the Kappa coefficient is 0.77, which is higher than the random forest and support vector machine methods in terms of classification accuracy.By adjusting and optimizing the decision tree model for soybean planting area identification and area estimation in the year without sample, the soybean distribution change in Heilongjiang province from 2017 to 2021 was obtained with a total accuracy of more than 90%, a Kappa coefficient of more than 0.60, and an accuracy of more than 95% for the area estimation results.

Key words: Vegetation index, Water index, Spectral analysis, Decision tree classification

中图分类号: