해양 어족자원의 실시간 모니터링을 위한 무선 수중 카메라 및 인공지능 기반 인식 시스템의 개발

Abstract

This paper proposes a system to monitor fishery resources remotely in real-time. A wireless monitoring system was designed to capture underwater videos using a fish-eye camera and send them to a server via Long Range and Long Term Evolution communication. A convolutional neural network then detects and classifies fish from the captured videos. The designed system was installed on four coastal fish farms and recognized the two species of fish accurately.