Ultra-Fast Segmentation And Quantification Of Polymetallic Nodule Coverage In High-Resolution Digital Images

The exploration of the seafloor regarding the abundances of poly-metallic nodules has received growing attention recently. Image-based exploration using cameraequipped deep-sea observation systems has been introduced successfully to increase spatial resolution in the exploration process but the evaluation and interpretation of the huge amount of image data creates a serious bottleneck. In this paper we present a new software solution to the problem of automatic detection and segmentation of poly-metallic nodules in underwater images, which is not only accurate but so fast that real time image analysis is definitely in reach. Thus, an application on an AUV (Autonomous Underwater Vehicle) seems possible in the near future.