Enhancing AI detection of dugong and other marine megafauna species (Old ID 30032)
Role
Principal Investigator
Description
Standardised aerial surveys have been conducted across northern Australia for over three decades to monitor dugong populations. JCU is currently monitoring dugongs across the entire eastern Queensland coast using conventional observer survey approach. In parallel to the monitoring work, JCU is experimenting the use of aerial images to conduct these large scales surveys. Preliminary results from the image processing work reveals that substantial efforts need to be put in to streamline and fast-track the processing of large image datasets to make imagery survey a cost-effective approach in the future. An AI for detecting dugongs is available but was developed based on images collected in Western Australian waters, a different habitat compared to eastern Queensland inshore waters. Preliminary tests ran by our team suggest that the current AI requires additional research work to improve its level of precision while other competitive AIs also need to be investigated. Upon completion of our tests, our team will outline steps toward the improvement of the current AI and/or exploration of alternative methods with the end goal of producing an automated approach that fast-tracks the processing of large image datasets collected during large-scale marine wildlife surveys.
Date
15 Jun 2023 - 01 May 2024
Project Type
CONTRACT_RESEARCH
Keywords
Dugong (Dugong dugon);Aerial Surveys;Conservation and Management;Aerial Imagery;Artificial Intelligence
Funding Body
Department of Climate Change, Environment, Energy and Water (DCCEEW)
Amount
45454.55
Project Team
Mohammad Jahanbakht