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
- 자료유형
- 학위논문
- 저자정보
- 지도교수
- 류관희
- 발행연도
- 2016
- 저작권
- 충북대학교 논문은 저작권에 의해 보호받습니다.
이용수1
이 논문의 연구 히스토리 (2)
- 이 논문의 후속연구가 궁금하신가요?
- 연관 학술논문 또는 학술발표를 통해 보다 발전된 연구결과를 확인하실 수 있습니다.
- 이 논문의 연구 히스토리 확인하기
초록· 키워드
오류제보하기
Recently, image processing technology has been used in a number of area, the trend in research has been made a lot due to an increase in the CCTV. Object detection and tracking is an essential element in the image processing. It has been used in a method of image reconstruction, wandering population tracking, vehicle license plate recognition, etc. However it is difficult not handle noise in the image due to suddenly changes in illumination, dynamic backgrounds and shadow.
In this paper, it introduces ROI based object extracting method using image feature to complement the above-mentioned problems. The method is captured images using Kinet. The purpose of this method is extract the silhouettes of moving object and solve the various problems of the image processing. We propose a detection method for moving objects in five steps. We first propose to subtract a background in depth camera which leads to discover the object’s location in the image. We then perform labeling and contouring in the image. Further, we select bounding box of a person and discover feature points in selected region. Lastly, using feature points, we construct a convex hull and detect the moving object. And convex hull points are adding the feature points. So we made a new contour of moving object. The results of simulations, which were carried out by a user who sits on a surfing board, demonstrate that the proposed method can adequately detect the user.
In this paper, it introduces ROI based object extracting method using image feature to complement the above-mentioned problems. The method is captured images using Kinet. The purpose of this method is extract the silhouettes of moving object and solve the various problems of the image processing. We propose a detection method for moving objects in five steps. We first propose to subtract a background in depth camera which leads to discover the object’s location in the image. We then perform labeling and contouring in the image. Further, we select bounding box of a person and discover feature points in selected region. Lastly, using feature points, we construct a convex hull and detect the moving object. And convex hull points are adding the feature points. So we made a new contour of moving object. The results of simulations, which were carried out by a user who sits on a surfing board, demonstrate that the proposed method can adequately detect the user.
목차
Ⅰ. 서 론 11.1 연구의 필요성 및 목적 11.2 연구의 내용 3Ⅱ. 관 련 연 구 42.1 키넥트의 User Index를 이용한 사용자 추출 42.2 영상 보행자 검출 기술 62.3 배경분리를 이용한 객체 검출 8Ⅲ. 영상 특징을 이용한 ROI 기반 객체 추출 기법 123.1 시스템 흐름도 133.2 객체 ROI 구하기 143.3 객체의 특징점 찾기 173.4 객체의 실루엣 추출 19Ⅳ. 실험 결과 224.1 구현결과 224.2 구현 성능 비교 23Ⅳ. 결 론 31참고문헌 33
최근 본 자료
댓글(0)
0