Detection of trawling patterns in seabed images

A large number of studies have proved that bottom trawling have great impact on benthic habitats. It is desirable to monitor the damaged level of seafloor and the relationships with ecosystem. Recent years, underwater video imagery has been shown a valuable approach to provide detailed information of physical properties of seabed. In this paper, a method to recognizing trawling marks in seabed image was presented. Instead of using techniques of texture process for seabed analysis, geometric feature-based methods have been applied for trawling pattern classification. The proposed method has been tested with real world images and an overall 83% recognition rate has been achieved with leave one out validation method. This offers the potential of helping scientist to process the amount of underwater images for fishery and seabed environment monitoring.