Scholarlink

Résumé Smart agriculture is a key sector that promotes the automation of agricultural systems and enhances crop monitoring. Numerous scientific studies rely on remote sensing indices to efficiently monitor vegetation. These radiometric indices, calculated from reflectance values measured by sensors onboard satellites or drones, are essential for monitoring agricultural ecosystems. In this study, we propose a vegetation index prediction method based on six indices: NDVI(Normalized Difference Vegetation Index), EVI(Enhanced Vegetation Index), NDWI(Normalized Difference Water Index), MNDWI (Modified NDWI), CVI(Chlorophyll Vegetation Index), and WBI(Water Body Index). To achieve better performance, we apply five machine learning models (CNN-LSTM, N-BEATS, RNN, Transformer, and Random Forest) to time series derived from Sentinel-2 images (2018–2023). The evaluation is based on comparing the mean squared error (MSE) and the mean absolute error (MAE). The fully connected N-BEATS model proves to be the most effective, achieving an MSE of approximately 0.005 and an MAE of 0.06, thus demonstrating its usefulness for proactive vegetation and agricultural resource management.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting the Health of Vegetation, Chlorophyll, and Water: The Case of the Groundnut Basin in Senegal with Machine Learning Models (HVIP)

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Pape El Hadji Abdoulaye Gueye,
  • Cherif Bachir Deme,
  • Adrien Basse

摘要

Résumé Smart agriculture is a key sector that promotes the automation of agricultural systems and enhances crop monitoring. Numerous scientific studies rely on remote sensing indices to efficiently monitor vegetation. These radiometric indices, calculated from reflectance values measured by sensors onboard satellites or drones, are essential for monitoring agricultural ecosystems. In this study, we propose a vegetation index prediction method based on six indices: NDVI(Normalized Difference Vegetation Index), EVI(Enhanced Vegetation Index), NDWI(Normalized Difference Water Index), MNDWI (Modified NDWI), CVI(Chlorophyll Vegetation Index), and WBI(Water Body Index). To achieve better performance, we apply five machine learning models (CNN-LSTM, N-BEATS, RNN, Transformer, and Random Forest) to time series derived from Sentinel-2 images (2018–2023). The evaluation is based on comparing the mean squared error (MSE) and the mean absolute error (MAE). The fully connected N-BEATS model proves to be the most effective, achieving an MSE of approximately 0.005 and an MAE of 0.06, thus demonstrating its usefulness for proactive vegetation and agricultural resource management.