AI-systems for marine debris

Developing machine vision for marine litter detection

Marine debris originating from human activity has been accumulating in underwater environments such as oceans, lakes, and rivers for decades. The extent, type, and amount of waste is spread over wast areas and is also largely unknown. Methods for detecting and mapping debris is therefore vital in order to effectively plan and execute physical removal. Skarv Technologies has developed a prototype "marine debris detection system" (MDDS) system using an autonomous underwater vehicle (AUV), equipped with an underwater hyperspectral imager (UHI) and stereo-camera. The system was tested in the city fjord Store Lungegårdsvann in Bergen, Norway with good results. Read more about the survey in our paper (Link).

Skarv Technologies marine debris system output

The system uses convolutional neural networks (CNNs) to match image patterns with known debris morphology. As the variability in marine trash is vast, the system will give a high amount of false poisitives. This problem is handled by subsequent post-processing which intelligently sorts the detections using supporting features coming from featured-based CNNs.