研究目的
To estimate missing spectral features, specifically the normalized difference vegetation index (NDVI), through data fusion and deep learning, exploiting both temporal and cross-sensor dependencies on Sentinel-1 and Sentinel-2 time-series.
研究成果
The proposed CNN-based data fusion approach shows promising results in estimating NDVI from multitemporal SAR images, highlighting the strong relationship between SAR and NDVI that can be captured through deep learning. The method is general and can be extended to other features and practical real-world applications.
研究不足
The approach requires a large amount of labelled data for training, and there can be large variations in performance from one date to another due to the variable correlation degree between NDVI and SAR along the seasons.
