Description:
Maritime environments present unique challenges for reliable perception and object detection, particularly in harsh and constrained conditions. This research addresses these challenges by developing advanced semantic segmentation and object detection techniques tailored for difficult maritime settings. Using multi-sensor fusion, the work enhances mapping and localization accuracy, ensuring robust navigation in complex environments. Deep neural networks are employed to bridge the gap between simulation and reality, utilizing sim-to-real data generation for improved training. These innovations aim to enhance the reliability and robustness of perception systems, enabling safer and more efficient autonomous operations in challenging maritime scenarios.
systems in support of ocean exploration