인문학
사회과학
자연과학
공학
의약학
농수해양학
예술체육학
복합학
지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 저자정보
초록·키워드
This paper introduces a new concept that involves YOLO object detection that will help apartment resicents separate the recyclables. The YOLO object detection will detect the partially covered recyclables help by the resicents and will automatically produce a notification to throw them into the correct bin accordingly by turning on a buzzer and LED attached to the corresponding bin. The data set for the object detection is gathered from multiple open-source platforms such as Kaggle that trained in YOLOv5 to classify four different types of recyclables, which is Glass, Can, Paper, and Plastic. Later on, the trained model will be run inside a microcomputer such as Raspberry Pi that will be installed on the trash bin around the neighbourhood. A simulation test is conducted by running the model on a Windows Pc with several recyclables brough to the front of the webcam one by one to trigger the object detection. On the simulation, the trained model resulted in a score of 0.992 on precision, 0.989 on recall, 0.991 on mAP@0.5, and 0.896@0.5:0.95. The result shows a promising number that provides enough headroom for a less powerful computer to still perform well without sacrificing too much accuracy.
본문·목차
인공지능 문자 인식 모델을 통해 추출된 텍스트로, 일부 오타나 오류가 포함될 수 있으나 지속적으로 개선 중입니다.
오류를 발견하셨다면 해당 부분을 드래그한 후 ' 를 통해 신고해주세요.
오류를 발견하셨다면 해당 부분을 드래그한 후 ' 를 통해 신고해주세요.