http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
심층 컨볼루션 신경망을 이용한 실시간 어획 어종 인식 및 카운팅 알고리즘 개발 및 시스템 구현
김승규(Seung-Gyu Kim),박세용(Se-Yong Park),송영남(Young-Nam Song),황신혁(Shin-Hyuk Hwang),임태호(Tae-Ho Im) 한국정보통신학회 2024 한국정보통신학회논문지 Vol.28 No.2
Overfishing, marine pollution, and climate change are exacerbating the global depletion of fishery resources, which has become a significant problem. To manage this issue and promote sustainable fishing, many countries are implementing the Total Allowable Catch (TAC) system. To this end, numerous countries, including the United States, Canada, and the EU, have introduced the Electronic Monitoring (EM) system, which requires fishing vessels to record their operations and submit the footage to the institutions responsible for managing fishery resources. In this paper, we researched a system that measures the catch volume from videos filmed in real-time on fishing vessels by integrating a CCTV system (EM) and deep learning technology. Using the deep convolutional neural network of the video data filmed in real-time by the CCTV system, we implemented and evaluated the recognition of catch species and a counting algorithm based on Nvidias Jetson hardware system to measure the catch volume.