메뉴 건너뛰기
소속 기관 / 학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(충북대학교) (충북대학교)
저널정보
제어로봇시스템학회 제어로봇시스템학회 논문지 제어로봇시스템학회 논문지 제26권 제7호
오류 신고하기

피인용 0

검색

    초록·키워드

    In autonomous vehicles, lidar and camera are used for object recognition. Because cameras are difficult to recognize at night, there is a lot of research on object recognition using lidar. Generally, it uses high channel lidar for object recognition, but it uses low channel lidar because it is expensive and difficult to use. Object recognition methods using low channel lidar are usually converted to 2D images. However, in the top-view method, the object recognition rate is low due to the low density, and the polar-view method has a disadvantage in that the recognition rate decreases in a complex environment such as a background and a building behind the object. To solve this problem, we propose an object recognition method that uses lidar data directly. First, the convolution is performed for each channel of Lidar data, and then the lidar data is integrated again. Then, a second convolution is performed using the height data. Finally, find the object in the detection network.

    최근 본 자료 전체보기

      UCI(KEPA) : I410-ECN-0101-2020-003-000879162