Fine extraction of multi-crop planting area based on deep learning with Sentinel- 2 time-series data

IF 5.8 3区 环境科学与生态学 0 ENVIRONMENTAL SCIENCES
Jingmin Jiang, Jiahua Zhang, Xue Wang, Shichao Zhang, Delong Kong, Xiaopeng Wang, Shawkat Ali, Hidayat Ullah
{"title":"Fine extraction of multi-crop planting area based on deep learning with Sentinel- 2 time-series data","authors":"Jingmin Jiang,&nbsp;Jiahua Zhang,&nbsp;Xue Wang,&nbsp;Shichao Zhang,&nbsp;Delong Kong,&nbsp;Xiaopeng Wang,&nbsp;Shawkat Ali,&nbsp;Hidayat Ullah","doi":"10.1007/s11356-025-36405-4","DOIUrl":null,"url":null,"abstract":"<div><p>Accurate and timely access to the spatial distribution of crops is crucial for sustainable agricultural development and food security. However, extracting multi-crop areas based on high-resolution time-series data and deep learning still faces challenges. Therefore, this study aims to provide an effective model for multi-crop classification using high-resolution remote sensing time-series data. We designed two deep learning models based on convolutional neural network-long short-term memory (CNN-LSTM) and bidirectional long short-term memory (Bi-LSTM). The monthly synthetic time series of the normalized difference vegetation index (NDVI) from Sentinel-2 data will be used as input features to extract the multi-crop planting area in Shandong province’s northwestern, southwestern, and eastern regions. The results showed that deep learning models achieved higher accuracy compared to the random forest (RF) and extreme gradient boosting (XGBoost) models, with CNN-LSTM achieving the highest overall accuracy of 96.48%. At the county level, the coefficients of determination (<i>R</i><sup>2</sup>) for the CNN-LSTM model were 0.91 for wheat, 0.88 for maize, and 0.73 for spring cotton. This study demonstrates that the CNN-LSTM model combined with monthly synthetic time-series NDVI provides a feasible approach for accurately mapping high-resolution multi-crop planting areas and also contributes significantly to decision support and resource management in agricultural production. </p></div>","PeriodicalId":545,"journal":{"name":"Environmental Science and Pollution Research","volume":"32 19","pages":"11931 - 11949"},"PeriodicalIF":5.8000,"publicationDate":"2025-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Environmental Science and Pollution Research","FirstCategoryId":"93","ListUrlMain":"https://link.springer.com/article/10.1007/s11356-025-36405-4","RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"0","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0

Abstract

Accurate and timely access to the spatial distribution of crops is crucial for sustainable agricultural development and food security. However, extracting multi-crop areas based on high-resolution time-series data and deep learning still faces challenges. Therefore, this study aims to provide an effective model for multi-crop classification using high-resolution remote sensing time-series data. We designed two deep learning models based on convolutional neural network-long short-term memory (CNN-LSTM) and bidirectional long short-term memory (Bi-LSTM). The monthly synthetic time series of the normalized difference vegetation index (NDVI) from Sentinel-2 data will be used as input features to extract the multi-crop planting area in Shandong province’s northwestern, southwestern, and eastern regions. The results showed that deep learning models achieved higher accuracy compared to the random forest (RF) and extreme gradient boosting (XGBoost) models, with CNN-LSTM achieving the highest overall accuracy of 96.48%. At the county level, the coefficients of determination (R2) for the CNN-LSTM model were 0.91 for wheat, 0.88 for maize, and 0.73 for spring cotton. This study demonstrates that the CNN-LSTM model combined with monthly synthetic time-series NDVI provides a feasible approach for accurately mapping high-resolution multi-crop planting areas and also contributes significantly to decision support and resource management in agricultural production. 

基于深度学习的Sentinel- 2时间序列多作物种植面积精细提取
准确及时地获取作物的空间分布对可持续农业发展和粮食安全至关重要。然而,基于高分辨率时间序列数据和深度学习的多作物区域提取仍然面临挑战。因此,本研究旨在为利用高分辨率遥感时间序列数据进行多作物分类提供一个有效的模型。我们设计了两个基于卷积神经网络的深度学习模型——长短期记忆(CNN-LSTM)和双向长短期记忆(Bi-LSTM)。利用Sentinel-2数据每月合成的归一化植被指数(NDVI)时间序列作为输入特征提取山东省西北、西南和东部地区的多作物种植面积。结果表明,与随机森林(RF)和极端梯度增强(XGBoost)模型相比,深度学习模型的准确率更高,其中CNN-LSTM的总体准确率最高,为96.48%。在县一级,CNN-LSTM模型的决定系数(R2)分别为小麦0.91、玉米0.88和春棉0.73。研究表明,CNN-LSTM模型结合月合成时间序列NDVI为高精度多作物种植区测绘提供了可行的方法,对农业生产中的决策支持和资源管理也有重要贡献。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
CiteScore
8.70
自引率
17.20%
发文量
6549
审稿时长
3.8 months
期刊介绍: Environmental Science and Pollution Research (ESPR) serves the international community in all areas of Environmental Science and related subjects with emphasis on chemical compounds. This includes: - Terrestrial Biology and Ecology - Aquatic Biology and Ecology - Atmospheric Chemistry - Environmental Microbiology/Biobased Energy Sources - Phytoremediation and Ecosystem Restoration - Environmental Analyses and Monitoring - Assessment of Risks and Interactions of Pollutants in the Environment - Conservation Biology and Sustainable Agriculture - Impact of Chemicals/Pollutants on Human and Animal Health It reports from a broad interdisciplinary outlook.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信