Development of Remote Underwater Monitoring System for Image Acquisition and Fish Detection using Neural Network

Abstract

This paper proposes a system that can remotely acquire and analyze underwater images in real-time. The proposed system utilizes Long Range communication for remote control, Long Term Evolution communication for realtime image transmission, and a neural network (NN) for underwater image analysis. As a result, the proposed system took less cost to install and operated for a long time. Two systems have been installed on an aquafarm and obtained underwater images continuously, and the NN could classify two fish species from acquired images. The proposed system can be applied for broader oceanic monitoring by attaching additional sensors and scaling up to acquire more data.