동적 환경 적응형 모델기반 카메라 추적
Model-Based Camera Tracking Adaptive to Dynamic Environments
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
모델기반 카메라 추적은 마커를 사용하지 않는 방법이지만, 강건성, 유연성을 보장 하기 때문에 많은 주목을 받아왔다. 대상 객체 혹은 장면의 정확한 3차원 그래픽 모델 및 대략적인 카메라 포즈가 주어졌을 때, 대략적인 카메라 포즈에 의해 투영된 모델을 카메라 영상에서 추출된 대상 객체 혹은 장면의 에지, 특징점과 같은 시각적 단서와 매칭시킴으로써 대략적인 카메라 포즈를 정확한 카메라 포즈로 갱신한다. 그러므로 카메라 영상은 많은 수의 특징적인 시각적 단서를 포함해야 하고, 대략적인 카메라 포 즈는 실제 (대상) 카메라 포즈와 유사해야 한다. 그러나 이러한 요구사항은 실 환경에 서는 만족되기 어렵다. 더구나 대부분의 모델기반 카메라 추적 방법들은 정적인 장면 을 가정하지만, 실 환경에서 장면은 동적으로 변한다. 즉, 조명 환경이 자주 변하고, 객체들이 움직이거나 가려지며, 새로운 객체가 나타나기도 한다. 그러므로 이러한 동 적인 환경에 적응적인 모델기반 카메라 추적 방법이 요구되는데, 이러한 방법을 제안 하는 것이 본 논문 내용의 핵심이다. 이를 위해 관련 요소 기술을 제안하고, 실 영상 및 합성 영상을 이용한 실험을 통해 제안된 방법을 정량적으로 평가한다.
Abstract
As a type of marker-less camera tracking, model-based camera tracking has received considerable attention due to its robustness and flexibility. Given an accurate and precise 3D graphic model of a target object/scene and a rough camera pose, it updates the rough camera pose to an accurate camera pose in such a way that the model projected by the rough camera pose is aligned in the camera image by tracking visual cues such as edges and feature points. For accuracy and reliability, therefore, the camera images must include a number of reliable and distinctive visual cues and the rough camera pose must be similar to the true (target) camera pose. However, these requirements are hard to satisfy in real-world situations. Furthermore, although most of existing model-based methods assume that the scene is static, the scene is not static in a real-world. That is, lighting condition is often changed, objects are moving/occluded, and new objects show up, etc. Therefore, a model-based camera tracking method being adaptive to such dynamic real-world situations is strongly desired, which is our concern in this paper. We propose several component techniques required for conventional model-based camera tracking to be adaptive to dynamic environments. The techniques are quantitatively evaluated with real and synthetic images.
참고문헌 (19)
- Bay H, "SURF: Speeded up robust features", Proc. of Euro, Conf. on Computer Vision, pp.404~417 (2006)
- Dornaika F, "Pose estimation using point and line correspondences", Real-Time Imaging, 5, pp.215~230 (1999)
- Drummond T, "Real-time visual tracking of complex structures", IEEE Trans, on Pattern Analysis and Machine Intelligence, 24(7), pp.932~946 (2002)
- Gouet BLV, "SAP: A robust approach to track objects in video streams with snakes and points", Proc. of British Machine Vision Conf, pp.737~746 (2004)
- Hartley R, "Multiple View Geometry" (2003), Cambridge University Press
- Kato H, "Marker tracking and HMD calibration for a video-based augmented reality conferencing system", Proc. of Intl. Work. on Augmented Reality, pp.85~94 (1999)
- Kiryati N, "A probabilistic Hough transform", Pattern Recognition, 24(4), pp.303~316 (1991)
- Lepetit V, "Point matching as a classification problem for fast and robust object pose estimation", Proc. of IEEE Conf. on Computer Vision and Pattern Recognition, pp.244~250 (2004)
- Lowe DG, "Robust model-based motion tracking through the integration of search and estimation", Intl. J. of Computer Vision, 8(2), pp.113~122 (1992)
- Lowe DG, "Distinctive image features from scale-invariant keypoints", Intl. J. of Computer Vision, 60(2), pp.91~110 (2004)
- Park H, "Object-adaptive tracking for AR guidance system", Proc. of Intl. Conf. on Virtual-Reality Continuum and Its Applications in Industry (2008)
- Park H, "Analytic fusion of visual cues in model-based camera tracking", Proc. of Intl. Conf. on Virtual-Reality Continuum and Its Applications in Industry, pp.215~220 (2009)
- Pressigout M, "Real-time 3D model-based tracking: Combining edge and texture information", Proc. of Intl. Conf. on Robotics and Automation, pp.2726~2731 (2006)
- Rosten E, "Fusing points and lines for high performance tracking", Proc. of Intl. Conf. on Computer Vision, 통권 : 2, pp.1508~1511 (2005)
- Shahrokni A, "Texture boundary detection for real-time tracking", Proc. of Euro. Conf. on Computer Vision (2004)
- Skrypnyk I, "Scene modeling, recognition and tracking with invariant image features", Proc. of IEEE and ACM Intl. Sym. on Mixed and Augmented Reality, pp.110~119 (2004)
- Snyder JM, "Ray tracing complex models containing surface tessellations", ACM SIGGRAPH Computer Graphics, 21(4), pp.119~128 (1987)
- Tran H, "Real-time keypoints matching: application to visual servoing", Proc. of Intl. Conf. on Robotics and Automation, pp.3787~3792 (2007)
- Vacchetti L, "Combining edge and texture information for real-time accurate 3D camera tracking", Proc. of IEEE and ACM Intl. Sym. on Mixed and Augmented Reality, pp.48~57 (2004)