今天给大家分享的论文是德国航空航天中心的遥感数据中心在remote sensing 期刊上发表的《Cropland and Crop Type Classification with Sentinel-1 andSentinel-2 Time Series Using Google Earth Engine forAgricultural Monitoring in Ethiopia》。本文利用 Sentinel-1 和 Sentinel-2 的时序遥感数据，通过 Google Earth Engine 平台，提出了一种针对埃塞俄比亚农田与作物类型分类的方法。研究分析了多种输入参数的分类效果，并验证了不同模型在三个区域和两年数据的适用性。
论文信息
题目：Cropland and Crop Type Classification with Sentinel-1 andSentinel-2 Time Series Using Google Earth Engine forAgricultural Monitoring in Ethiopia
关键词：cropland; crop types; Ethiopia;Google Earth Engine; LULC; multispectral data; radardata; random forest classification; Sentinel-1; Sentinel-2; time series
作者：Christina Eisfelder, Bruno Boemke, Ursula Gessner, Patrick Sogno,Genanaw Alemu, Rahel Hailu, Christian Mesmer, Juliane Huth
DOI：10.3390/rs16050866
论文摘要
Cropland monitoring is important for ensuring food security in the context of global climatechange and population growth.Freely available satellite data allow for the monitoring of large areas,while cloud-processing platforms enable a wide user community to apply remote sensing techniques.Remote sensing-based estimates of cropped area and crop types can thus assist sustainable landmanagement in developing countries such as Ethiopia.In this study,we developed a method forcropland and crop type classification based on Sentinel-1 and Sentinel-2 time-series data using GoogleEarth Engine.Field data on 18 different crop types from three study areas in Ethiopia were availableas reference for the years 2021 and 2022.First,a land use/land cover classification was performedto identify cropland areas.We then evaluated different input parameters derived from Sentinel-2and Sentinel-1,and combinations thereof,for crop type classification.We assessed the accuracyand robustness of 33 supervised random forest models for classifying crop types for three studyareas and two years.Our results showed that classification accuracies were highest when Sentinel-2spectral bands were included.The addition of Sentinel-1 parameters only slightly improved theaccuracy compared to Sentinel-2 parameters alone.The variant including S2 bands,EVI2,and NDRe2from Sentinel-2 and VV,VH,and Diff from Sentinel-1 was finally applied for crop type classification.Investigation results of class-specific accuracies reinforced the importance of sufficient referencesample availability.The developed methods and classification results can assist regional experts inEthiopia to support agricultural monitoring and land management.
耕地监测在全球气候变化和人口增长背景下对确保粮食安全至关重要。免费提供的卫星数据使得对大范围地区进行监测成为可能，而云端处理平台则使广泛的用户群体能够应用遥感技术。基于遥感的耕地面积和作物类型估算可以帮助发展中国家（如埃塞俄比亚）实现可持续土地管理。本研究开利用Google Earth Engine平台发了一种基于Sentinel-1和Sentinel-2时间序列数据的耕地和作物类型分类方法。以2021年和2022年埃塞俄比亚三个研究区的18种不同作物类型的实地数据作为参考数据。首先，研究进行土地覆盖分类以识别耕地区域。然后评估了从Sentinel-2和Sentinel-1及其组合中提取的不同输入参数，以进行作物类型分类。研究评估了33个监督随机森林模型在三个研究区和两年期间对作物类型分类的准确性和鲁棒性。研究结果表明，当包含Sentinel-2光谱波段时，分类精度最高。与仅使用Sentinel-2参数相比，加入Sentinel-1参数对分类精度的提升相对较小。最后，结合了Sentinel-2波段、EVI2和NDRe2，以及Sentinel-1的VV、VH和差分波段用于作物类型分类。对类别特定精度的调查结果进一步验证了参考样本数量的重要性。研究所开发的方法和分类结果可以帮助埃塞俄比亚的地区专家支持农业监测和土地管理。
重要图表
——研究区概况图——
—— 技术流程图——
——三个研究区在2021年和2022年野外活动期间收集的野外数据点的位置——
——三个研究区土地利用覆盖分类——
——对输入数据集的33个变量进行作物分类的总体精度比较——
——3个研究区2021年和2022年的作物类型分类图——
——用于作物类型分类的参考数据数量以及具体类别的（a）生产者准确率、（b）用户准确率和（c）F1 分数——
本文创新
该研究系统分析了不同 Sentinel 数据组合对分类精度的影响，提出了一种优化作物分类的模型。
心得收获
针对作物分类的优化，需要充分考虑输入数据的特性及研究区的具体情况，例如农田分布和作物生长季节。
欢迎指出文中翻译存在不准确的地方！
