인문학
사회과학
자연과학
공학
의약학
농수해양학
예술체육학
복합학
DBpia AI
AI 에이전트
AI 아이디어
AI 리더
AI 뷰어
Citeasy
크롬 서지관리 다운로드
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.
학술대회자료
Full-text
오류 신고하기해당 페이지 내 제목·저자·목차·페이지정보가 잘못된 경우 알려주세요!
초록·키워드
Vehicle detection using a camera has been studied actively for a couple of decades in the intelligent vehicle system. Especially, vehicle detection at nighttime has several difficulties because of lack of light. Most of night vehicle detection methods use a high dynamic range (HDR) camera or tune brightness of a camera to capture only light regions. However, the camera we use does not have HDR property and is tuned for lane departure warning system (LDWS). Therefore, the brightness of the camera is fixed to capture lanes on the road even at night and the white balance of the camera is changed to boost white and yellow. In this paper, we propose an effective vehicle detection method using non-HDR camera. The proposed method consists of two main parts; light segmentation and pairing. In the light segmentation part, candidate lights are segmented from an original image using a novel adaptive threshold method. Features for a classifier are extracted from each candidate light. Candidate lights are classified as tail, head, and other lights using a random forest classifier. In the pairing part, only tail lights are handled to detect preceding vehicles. Two tail lights whose Y axis position difference is smaller than threshold are collected as a pairing candidate. Features for pairing classification are extracted from the pairing candidates. Random forest classifier is also used to classify pairing candidates as a vehicle or a non-vehicle. Experiments show that the proposed method effectively detects vehicles in the several different environments.
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.
최근 본 자료 전체보기
UCI(KEPA) : I410-ECN-0101-2013-556-001415227
