Active learning applied to marine geo-physical data for offshore-wind site selection: Detecting and sizing seabed boulders in sonar imagery

Details

Supervisors

Macq, Benoît

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil en mathématiques appliquées, à finalité spécialisée

Abstract

enThe aim of this master’s thesis is to investigate how active learning techniques can be applied to reduce the manual effort required for annotating sonar imagery of underwater environments. The primary objective is to improve the efficiency of machine learning models in analyzing boulders captured by side-scan sonar, by selecting the most informative data samples to label. This includes both the identification of boulders within sonar images and the estimation of their physical dimensions. The broader goal is to enhance the scalability of automated seabed analysis by minimizing human involvement while maintaining high prediction accuracy.