Exploring the Deep: Machine Learning and Remote Sensing for Targeted Underwater Exploration and Mapping

Building directly on this foundation, new research is focused on fusing very high-resolution 3D bathymetric data with sidescan sonar for underwater target detection using deep learning under a range of environmental and optical conditions. The targets identified by the deep learning model will be ranked using a forest-based model that incorporates target morphometrics, and these results will be compared to human review. Looking ahead, this multi-modal approach will be scaled for broader seafloor environmental characterization. Future work aims to provide essential baseline data for the understudied Ivorian Continental Shelf, directly supporting efforts to establish crucial marine protected areas.

 

Bio
Dr. Leila Character is an Assistant Professor in the Department of Geography at Texas A&M University. She is a geospatial scientist working on solving hard, real-world environmental problems using remote sensing and machine learning. She has several projects focused on underwater mapping and object detection that seek to develop and implement approaches enabling the use of huge amounts of remotely sensed data to find and map features over large areas: either motivated because manual imagery analysis is far too time-consuming or because combining different types or layers of remotely sensed data and then processing using machine learning can produce new information, and entirely new GIS layers, that would be impossible to calculate or derive using just the human eye. These machine learning approaches enable insights that haven’t been possible in the past. Overall, her research is focused on three areas: machine learning model generalizability and methodological transferability, imagery preprocessing and data annotation, and data collection. She has experience working with many types of remotely sensed data, including hyper- and multi- spectral; RGB; lidar, radar, and sonar; and magnetometer. Her projects include both underwater and terrestrially based.

 

Dr. Character has a bachelor’s degree in Geology frow Sewanee: The University of the South. She completed a master’s degree and PhD in Geography at the University of Texas at Austin (UT Austin). Her postdoctoral fellowship at the University of Delaware in the School of Marine Science and Policy was focused on machine learning for underwater aircraft wreck detection using sidescan sonar. She also has professional experience as a geologist, environmental scientist, and at an artificial intelligence start-up.