Seafloor Image

Seafloor image analysis focuses on extracting meaningful information from underwater imagery, primarily to understand seabed composition, map bathymetry, and identify objects of interest. Current research heavily utilizes deep learning, employing convolutional neural networks (CNNs), vision transformers (ViTs), and autoencoders for tasks like classification, object detection, and semantic segmentation of seafloor images and sub-bottom acoustic data. These advancements are crucial for improving autonomous underwater vehicle (AUV) navigation, environmental monitoring (e.g., benthic habitat mapping, endangered species detection), and supporting various engineering projects requiring accurate seabed characterization. The development of large, publicly available datasets like BenthicNet is also significantly accelerating progress in this field.