Near Real-Time Flood Mapping with Weakly Supervised Machine Learning

• Published in: Remote sensing (Basel, Switzerland) Vol. 15; no. 13; p. 3263
• Main Authors: Vongkusolkit, Jirapa, Peng, Bo, Wu, Meiliu, Huang, Qunying, Andresen, Christian G.
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.07.2023
• Subjects: Algorithms; Annotations; Artificial neural networks; Classification; Climate change; Computer vision; Data mining; decision support systems; Decision trees; Deep learning; Edge detection; Emergency preparedness; flood detection; flood extent mapping; Flood mapping; Floods; humans; Hurricanes; Image processing; Labeling; Learning algorithms; Machine learning; Machine vision; Mapping; Methods; Multilayers; Neural networks; Pixels; Real time; Remote sensing; Supervised learning; United States; Urban areas; Urbanization; weakly supervised; United States

Advances in deep learning and computer vision are making significant contributions to flood mapping, particularly when integrated with remotely sensed data. Although existing supervised methods, especially deep convolutional neural networks, have proved to be effective, they require intensive manual...