Evaluation of Sentinel-1 & 2 time series to derive crop phenology and biomass of wheat and rapeseed: northen France and Brittany case studies

Mercier Audrey, Betbeder Julie, Baudry Jacques, Denize Julien, Leroux Vincent, Roger Jean-Luc, Spicher Fabien, Hubert-Moy Laurence. 2019. Evaluation of Sentinel-1 & 2 time series to derive crop phenology and biomass of wheat and rapeseed: northen France and Brittany case studies. In : Remote Sensing for Agriculture, Ecosystems, and Hydrology XXI. Neale Christopher M.U. (ed.), Maltese Antonino (ed). Strasbourg : SPIE, 20 p. (Proceedings of SPIE, 11149) SPIE Conference Remote Sensing for Agriculture, Ecosystems and Hydrology. 31, Strasbourg, France, 21 Octobre 2019/26 Octobre 2019.

Résumé : Crop monitoring at a fine scale is critical from an environmental perspective since it provide crucial information to combine increased food production and sustainable management of agricultural landscapes. The recent Synthetic Aperture Radar (SAR) Sentinel-1 (S-1) and optical Sentinel-2 (S-2) time series offer a great opportunity to monitor cropland (structure, biomass and phenology) due to their high spatial and temporal resolutions. In this study, we assessed the potential of Sentinel data to derive Wet Biomass (WB), Dry biomass (DB), water content and crop Phenological Stages (PS). This study focuses on wheat and rapeseed, which represent two of the most important seasonal crops of the world in terms of occupied area. Satellites and ground data were collected over two French temperate agricultural landscapes, in northern France and Brittany. Spectral bands and vegetation indices were derived from the S-2 images and backscattering coefficients and polarimetric indicators from the S-1 images. We used linear models to estimate the Crop Parameters (CP) of wheat and rapeseed crops. Satellite images were then classified using a random forest incremental procedure based on the importance rank of the input features to discriminate PS. Results showed that S-1 features were more efficient than S-2 features to estimate CP of rapeseed while S-2 features were better for wheat. We demonstrated the high potential of S-1 & 2 to predict principal PS (kappa=0.75) while secondary PS were misclassified. For wheat, the succession of PS predicted was consistent, further research is required to confirm the potential of S-1 & 2.

Mots-clés libres : Remote Sensing, Agriculture, Optical and SAR data

Auteurs et affiliations

• Mercier Audrey, CIRAD-ES-UPR Forêts et sociétés (FRA)
• Betbeder Julie, CIRAD-ES-UPR Forêts et sociétés (CRI)
• Baudry Jacques, INRA (FRA)
• Denize Julien, Université de Rennes 2 (FRA)
• Leroux Vincent, Université de Picardie (FRA)
• Roger Jean-Luc, INRA (FRA)
• Spicher Fabien, UPJV (FRA)
• Hubert-Moy Laurence, Université de Rennes 2 (FRA)

Source : Cirad-Agritrop (https://agritrop.cirad.fr/594481/)

Métriques

[ Page générée et mise en cache le 2024-03-29 ]