An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learning
It is imperative to rapidly and precisely acquire crop lodging area and severity for disaster prevention and yield prediction. However, estimation of crop lodging area at a large scale remains challenging due to the relatively low sensitivity of remote sensing signal to the lodging variation, limite...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-09-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843222001832 |
| _version_ | 1811183451385102336 |
|---|---|
| author | Haixiang Guan Jianxi Huang Xuecao Li Yelu Zeng Wei Su Yuyang Ma Jinwei Dong Quandi Niu Wei Wang |
| author_facet | Haixiang Guan Jianxi Huang Xuecao Li Yelu Zeng Wei Su Yuyang Ma Jinwei Dong Quandi Niu Wei Wang |
| author_sort | Haixiang Guan |
| collection | DOAJ |
| description | It is imperative to rapidly and precisely acquire crop lodging area and severity for disaster prevention and yield prediction. However, estimation of crop lodging area at a large scale remains challenging due to the relatively low sensitivity of remote sensing signal to the lodging variation, limited availability of remote sensing images, and lodging statistical data. This study proposes a new method for lodging area estimation based on the optimal grid cell of Sentinel-2 and crop lodging percentage, overcoming the limitation of traditional pixel-based mapping approaches that fail to obtain quantitative lodging information. Basing the spatial aggregation method, we analyzed the optimal grid size of Sentinel-2 data for lodging percentage estimation. Then we investigated the spectral response for different lodging percentage levels and analyzed the potential of lodging percentage estimation for Sentinel-2 metrics (including selected spectral bands and their derived vegetation indexes (VIs)). A quantitative model was established between the training set and the Sentinel-2 metrics using the random forest (RF) algorithm. Finally, around 1462.62 ha fields from six counties or districts in Heilongjiang province in China were estimated for lodging percentage. Results indicate that the proposed method can estimate the crop lodging percentage on the testing set with an R2 and RMSE of 0.64 and 25.24, respectively, which can explain around 95 % spatial variation of lodging crop. Moreover, the overall magnitude of reflectance increased with the increase in lodging percentage. Among all Sentinel-2 optimal metrics, the Green, SWIR1, and Red edge 1 bands are the most crucial indicators for lodging percentage estimation. Our results on lodging percentage estimation in the study area indicate that there is more lodging maize in the Meilisidawoerzu district than in other areas. Although typhoons passed over Fuyu and Lindian counties, the lodging percentage in these areas is relatively low. The lodging percentage map has great value in agriculture management and insurance claim. |
| first_indexed | 2024-04-11T09:46:18Z |
| format | Article |
| id | doaj.art-5b5ac59b0d1a43ea90eb781b9a4242c6 |
| institution | Directory Open Access Journal |
| issn | 1569-8432 |
| language | English |
| last_indexed | 2024-04-11T09:46:18Z |
| publishDate | 2022-09-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj.art-5b5ac59b0d1a43ea90eb781b9a4242c62022-12-22T04:30:57ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322022-09-01113102992An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learningHaixiang Guan0Jianxi Huang1Xuecao Li2Yelu Zeng3Wei Su4Yuyang Ma5Jinwei Dong6Quandi Niu7Wei Wang8College of Land Science and Technology, China Agricultural University, Beijing 100083, ChinaCollege of Land Science and Technology, China Agricultural University, Beijing 100083, China; Key Laboratory of Remote Sensing for Agri-Hazards, Ministry of Agriculture and Rural Affairs, Beijing 100083, China; Corresponding author at: College of Land Science and Technology, China Agricultural University, China.College of Land Science and Technology, China Agricultural University, Beijing 100083, China; Key Laboratory of Remote Sensing for Agri-Hazards, Ministry of Agriculture and Rural Affairs, Beijing 100083, ChinaCollege of Land Science and Technology, China Agricultural University, Beijing 100083, China; Key Laboratory of Remote Sensing for Agri-Hazards, Ministry of Agriculture and Rural Affairs, Beijing 100083, ChinaCollege of Land Science and Technology, China Agricultural University, Beijing 100083, China; Key Laboratory of Remote Sensing for Agri-Hazards, Ministry of Agriculture and Rural Affairs, Beijing 100083, ChinaSchool of Geography and Information Engineering, China University of Geosciences, Wuhan 430074, ChinaKey Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, ChinaCollege of Land Science and Technology, China Agricultural University, Beijing 100083, ChinaCollege of Land Science and Technology, China Agricultural University, Beijing 100083, ChinaIt is imperative to rapidly and precisely acquire crop lodging area and severity for disaster prevention and yield prediction. However, estimation of crop lodging area at a large scale remains challenging due to the relatively low sensitivity of remote sensing signal to the lodging variation, limited availability of remote sensing images, and lodging statistical data. This study proposes a new method for lodging area estimation based on the optimal grid cell of Sentinel-2 and crop lodging percentage, overcoming the limitation of traditional pixel-based mapping approaches that fail to obtain quantitative lodging information. Basing the spatial aggregation method, we analyzed the optimal grid size of Sentinel-2 data for lodging percentage estimation. Then we investigated the spectral response for different lodging percentage levels and analyzed the potential of lodging percentage estimation for Sentinel-2 metrics (including selected spectral bands and their derived vegetation indexes (VIs)). A quantitative model was established between the training set and the Sentinel-2 metrics using the random forest (RF) algorithm. Finally, around 1462.62 ha fields from six counties or districts in Heilongjiang province in China were estimated for lodging percentage. Results indicate that the proposed method can estimate the crop lodging percentage on the testing set with an R2 and RMSE of 0.64 and 25.24, respectively, which can explain around 95 % spatial variation of lodging crop. Moreover, the overall magnitude of reflectance increased with the increase in lodging percentage. Among all Sentinel-2 optimal metrics, the Green, SWIR1, and Red edge 1 bands are the most crucial indicators for lodging percentage estimation. Our results on lodging percentage estimation in the study area indicate that there is more lodging maize in the Meilisidawoerzu district than in other areas. Although typhoons passed over Fuyu and Lindian counties, the lodging percentage in these areas is relatively low. The lodging percentage map has great value in agriculture management and insurance claim.http://www.sciencedirect.com/science/article/pii/S1569843222001832Crop lodgingSpatial aggregationLodging percentage mapSpectral bandsVegetation indexesRegion scale |
| spellingShingle | Haixiang Guan Jianxi Huang Xuecao Li Yelu Zeng Wei Su Yuyang Ma Jinwei Dong Quandi Niu Wei Wang An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learning International Journal of Applied Earth Observations and Geoinformation Crop lodging Spatial aggregation Lodging percentage map Spectral bands Vegetation indexes Region scale |
| title | An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learning |
| title_full | An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learning |
| title_fullStr | An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learning |
| title_full_unstemmed | An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learning |
| title_short | An improved approach to estimating crop lodging percentage with Sentinel-2 imagery using machine learning |
| title_sort | improved approach to estimating crop lodging percentage with sentinel 2 imagery using machine learning |
| topic | Crop lodging Spatial aggregation Lodging percentage map Spectral bands Vegetation indexes Region scale |
| url | http://www.sciencedirect.com/science/article/pii/S1569843222001832 |
| work_keys_str_mv | AT haixiangguan animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT jianxihuang animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT xuecaoli animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT yeluzeng animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT weisu animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT yuyangma animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT jinweidong animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT quandiniu animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT weiwang animprovedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT haixiangguan improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT jianxihuang improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT xuecaoli improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT yeluzeng improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT weisu improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT yuyangma improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT jinweidong improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT quandiniu improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning AT weiwang improvedapproachtoestimatingcroplodgingpercentagewithsentinel2imageryusingmachinelearning |