Underwater Animal Detection Using YOLOV4

Abstract

Underwater computer vision system has been widely used for many underwater applications such as ocean exploration, biological research and monitoring underwater life sustainability. However, in counterpart of the underwater environment, there are several challenges arise such as water murkiness, dynamic background, low light and low visibility which limits the ability to explore this area. To overcome these challenges, there is a crucial to improve underwater vision system that able to efficiently adapt with varying environments. Therefore, it is great of significance to propose an efficient and precise underwater detection by using YOLOv4 based on deep learning algorithm. In the research, an open-source underwater dataset was used to investigate YOLOv4 performance based on metrics evaluation of precision and processing speed (FPS). The result shows that YOLOv4 able to achieve a remarkable of 97.96% for mean average precision with frame per second of 46.6. This study shows that YOLOv4 model is highly significant to be implemented in underwater vision system as it possesses ability to accurately detect underwater objects with haze and low-light environments.