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
- 박찬식
- 발행연도
- 2016
- 저작권
- 충북대학교 논문은 저작권에 의해 보호받습니다.
이용수3
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
In this paper, a visual simultaneous localization and mapping(SLAM) method which uses rectangle features is proposed in factor graph manner.
Using the rectangle reconstruction method called coupled line camera algorithm which calculates analytically the aspect ratio of the rectangle from perspectively warped quadrilateral in the image plane, the plane homography is derived. The stereo configuration of the camera is proposed in order to solve the ambiguity arising from we not know the true size of the rectangle without additional information with corresponding modification of the coupled line camera algorithm to reconstruct the rectangle even the scale. A relative pose between rectangle in the real world plane and the camera is calculated for each quadrilateral from normalized homography with the intrinsic parameters of the camera which is constant in images captured by the identically same camera. the performance of this pose estimation method is showed in the simulated environment with ground truth value.
Extracted feature and the relative position are consisted factor graph form and formulated in a nonlinear least-squares problem of the negative log likelyhood estimation of the graph. this problem is derived to the linear least squares problem through the linearization and solved by cholesky factorization for efficient calculation. And rectangle feature extraction methods are proposed by the line extraction based method and image segmentation based methods that is more effective approach than line extraction based method. the experiment was performed for the entire rectangle feature based SLAM system on KITTI Benchmark Suite that is widely used for analysis of the performance of the camera-based pose estimation system.
Using the rectangle reconstruction method called coupled line camera algorithm which calculates analytically the aspect ratio of the rectangle from perspectively warped quadrilateral in the image plane, the plane homography is derived. The stereo configuration of the camera is proposed in order to solve the ambiguity arising from we not know the true size of the rectangle without additional information with corresponding modification of the coupled line camera algorithm to reconstruct the rectangle even the scale. A relative pose between rectangle in the real world plane and the camera is calculated for each quadrilateral from normalized homography with the intrinsic parameters of the camera which is constant in images captured by the identically same camera. the performance of this pose estimation method is showed in the simulated environment with ground truth value.
Extracted feature and the relative position are consisted factor graph form and formulated in a nonlinear least-squares problem of the negative log likelyhood estimation of the graph. this problem is derived to the linear least squares problem through the linearization and solved by cholesky factorization for efficient calculation. And rectangle feature extraction methods are proposed by the line extraction based method and image segmentation based methods that is more effective approach than line extraction based method. the experiment was performed for the entire rectangle feature based SLAM system on KITTI Benchmark Suite that is widely used for analysis of the performance of the camera-based pose estimation system.
목차
1. 서론 11.1 연구 배경 11.2 연구 목적 및 제안 52. 사각형 특징의 기하학적 성질을 이용한 스테레오 카메라의 자세 복원 방법 92.1 평면 Homography를 이용한 카메라의 자세 복원 방법 92.2 선분 카메라쌍 방법을 이용한 사각형 복원 방법 112.3 3차원 자세 추정을 위한 스테레오 카메라 기반의 선분카메라쌍 방법 143. 사각형 특징을 이용한 Factor Graph 기반 Visual SLAM 213.1 SLAM을 위한 Factor Graph 표현 유도 213.2 최소자승법을 이용한 Factor Graph기반 위치 추정 및 지도 작성 244. 사각형 기반 Visual SLAM을 위한 특징 추출 방법 284.1 직선 추출 알고리즘을 응용한 사각형 특징 추출 284.2 영상 영역 분할 기반의 사각형 특징 추출 305. 실험결과 365.1 선분카메라쌍 기반의 자세 추정 성능 실험 365.2 사각형 특징 기반 factor graph SLAM 성능 실험 475.2.1 KITTI Dataset 실험 476. 결론 및 향후 연구 51
최근 본 자료
댓글(0)
0