Next day fire prediction via semantic segmentation

arXiv:2403.13545v1 Announce Type: cross
Abstract: In this paper we present a deep learning pipeline for next day fire prediction. The next day fire prediction task consists in learning models that receive as input the available information for an area up until a certain day, in order to predict the occurrence of fire for the next day. Starting from our previous problem formulation as a binary classification task on instances (daily snapshots of each area) represented by tabular feature vectors, we …

abstract arxiv cs.cv cs.lg deep learning fire information next paper pipeline prediction segmentation semantic type via