Research On Marine Creature Recognition From Seafloor Videos

Because of the complexity and diversity of marine creatures, there are tremendous difficulties in the study on ocean. The underwater camera deployed on these platforms has increased the capacity to study marine creatures. Many redundant data in these videos from the underwater camera result in excessive waste of storage resources and difficulties in valid information retrieval. This paper proposed a seafloor video processing method combining computer algorithm and public participation. This method consists of valid video clips extraction and an analysis of the species of marine creatures. Four videos were selected to verify this method, the results showed that the valid frames clips of four videos were extracted properly with the accuracies of higher than 80%, and all creatures in video were identified.