CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleA Study o Some Key Problems in 3D Reconstruction
Thesis Advisor胡占义
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Abstract作为计算机视觉领域中的重要研究方向之一,三维重建技术有着很高的理论研究价值 和广泛的应用前景。到目前为止,国内外在此领域取得了丰硕的研究成果。本论文在前人 的基础上,对文献中三维重建领域存在的一些主要问题进行了研究,所完成的工作主要有: (1)利用L M方法解决了运动重建中的捆绑调整问题,该方法在我们研究组的三维重建系 统实现中得到了很好的应用。 (2)系统研究了立体视觉中两副图像间的匹配问题。在短基线情况下,着重介绍了一种基于 极线的点匹配方法,并针对该方法提出了一些改进的措施;在长基线情况下,介绍了当 前主要的几种解决方法,并且就其中的一些方法给出了实验结果。 (3)直线提取是基于特征的模式识别的一个重要步骤。由于图像特征点的离散误差和特征点 提取过程中的位移现象,使得由二图像点所确定的直线的参数随之产生某种不确定性。 本文比较系统地研究了这种不确定性的分布,并讨论了这些理论结果在计算机视觉中的 应用。 (4)天文学中的星表匹配问题一直是天文观测中的核心问题之一,也是一个典型的模式识别 问题。本文利用RANSAC方法对星表的自动匹配问题进行了研究。实验结果表明该方 法能够快速有效地完成星表匹配的问题。
Other AbstractAs one of the most important research directions in computer vision, 3D reconstruction is of both theoretical and practical importance. Till now, the researchers have obtained fruitful achievements in this field. In this thesis, we investigated a number of important problems related to 3D reconstruction. The main work can be summarized as follows: (a) L_M algorithm was implemented to solve bundle adjustment problem in the field of structure from motion. It is showed the algorithm works well in our 3D reconstruction system. (b) We systematically investigated the stereo correspondence problem between two images. In the short baseline situation, a robust technique was elaborated for matching two uncalibrated images through the recovery of the unknown epipolar geometry and some new measures were also proposed to improve the matching results. In the wide baseline situation, some main methods in the literature are reviewed and some experimental results are provided. (c) Lines Extraction is an important process in feature based pattern recognition. Due to inevitable discretization error of feature points and location shift introduced in edge detection, the line determined by a pair of feature points gives rise to some uncertainty. We investigated the distribution of this uncertainty systematically and discussed its application in computer vision. (d) The problem of atlas matching is one of the important problems in astronomy. We used an RANSAC based technique to tackle the atlas-matching problem. The technique is proved to be robust, accurate, and computational efficient through extensive experiments.
Other Identifier630
Document Type学位论文
Recommended Citation
GB/T 7714
胡宇. 三维重建若干问题研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2002.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[胡宇]'s Articles
Baidu academic
Similar articles in Baidu academic
[胡宇]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[胡宇]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.