Forward-Looking Multibeam - Marine Litter Detection and Tracking Dataset (FLM-MLDT)

Institute for Systems and Computer Engineering, Technology and Science (INESC TEC) aims to make a significant contribution to the advancement of marine robotics and environmental monitoring by providing a novel dataset: the Forward-Looking Multibeam - Marine Litter Detection and Tracking Dataset (FLM-MLDT). Addressing the critical scarcity of labelled underwater data, this work presents 8,156 acoustic images acquired using the Kongsberg M3 High-Frequency Multibeam Echosounder with a forward-looking configuration installed in the PORTUS Unmanned Surface Vehicle (USV). Collected in the manoeuvring basin of Leixões harbour, this dataset offers information on 10 distinct types of marine debris. It provides input for developing deep learning algorithms, refining sonar signal processing, and enhancing the perception capabilities of different platforms for data extraction and environmental interaction. By making this dataset publicly available, INESC TEC seeks to drive further innovation in underwater computer vision and automated clean-up solutions, fostering a cleaner and more sustainable ocean environment.