AUTOMATED CLASSIFICATION OF CROP TYPES AND CONDITION IN A MEDITERRANEAN AREA USING A FINE-TUNED CONVOLUTIONAL NEURAL NETWORK

This study seeks to present the as-yet-unknown potential use of WorldView-3 for the mapping of methane point source emissions. The proposed retrieval methodology is based on the idea that the spectral channels not affected by methane can be used to predict the methane-affected band through regression analysis. The results show the precise location of 26 independent point emissions over different methane hotspot regions worldwide, which prove the game-changing potential that this mission entails.