CASIA OpenIR  > 毕业生  > 硕士学位论文
Thesis Advisor胡占义
Degree Grantor中国科学院自动化研究所
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
Keyword随机hough变换(Rht) 禁忌搜索(Ts) 纯随机方法(Ransac) 几何基元提取 参数空间分解法 优化方法 类星体识别 Randomized Hough Transform (Rht) Tabu Search (Ts) Random Sample Consensus (Ransac) Geometric Primitive Extraction Parameter Spa
Abstract几何基元提取是机器人视觉领域最根本的问题之一。它是任何计算机视觉 系统以至任何计算机视觉问题的关键组成部分和基本要求。Hough变换则是目 前应用最广的几何基元提取方法。然而随着理论研究的深入和Hough变换在实 际应用中的拓展,它的一些问题也暴露了出来:一个是Hough变换的空间消耗 大的问题:一个是Hough变换的计算时间长的问题,也即是算法效率问题。 为了解决这两个问题,人们做出了各种各样的努力。最近几年Hough变换 研究取得了较快的进展,文献中提出了许多改进的Hough变换方法,随机Hough 变换(RHT)即是其一。本文在RHT的基础上,对如何解决Hough变换所存 在的空间消耗大和计算效率低的问题做了进一步的研究,所做的主要工作如下: 1.针对Hough变换在高维基元提取中的应用,提出了参数空间分解法。 其基本思想是由多个二维数组来实现一个高维参数空间,从而大大降低了空间 开销。参数空间分解法实质上是在空间开销和噪声影响之间的一个均衡,是一 种以少量增加误识别率来换取大量减少空间消耗的策略,它使得Hough变换在 高维基元提取中的应用成为可能。大量实验证明参数空间分解法是一种有效的 Hough变换实现方法。 2.对几种常用的基元提取算法(随机Hough变换,禁忌搜索算法,纯随 机方法)做了算法比较。在提取单个基元所需对最小点集的采样次数的期望值 这一判别准则的基础上,对RHT和TS,RANSAC这三种方法从理论上作了较 为系统的分析,并得出了在几何基元提取中,Hough变换方法要更优于基于优 化算法的基元提取算法的结论。大量的模拟和实际图象实验也证明了这一点。 3.提出了一种基于随机Hough变换的三角形直接提取方法。大量模拟与 实际图象实验证实了该方法的有效性。 4.将Hough变换在类星体波形识别中进行了应用,在基于双参数的Hough变换的基础上提出了一种新的Hough变换波形识别方法,实验证明它具有简单、 快速、高效、鲁棒性强、通用性强等特点。
Other AbstractGeometric primitive extraction is one of the fundamental problems in the computer vision field. It plays a very important role in almost all the computer vision related problems. The Hough transform has been a widely used technique for geometric primitive extraction. However, two main problems embedded in the traditional Hough Transform greatly circumscribed its further applications. The one problem is its high memory requirement, the other is its heavy computational load. In order to cope with these two problems, many novations were proposed in the Hough field. The Randomized Hough Transform (RHT) was one of such novations, which is widely considered to be the most proper one. The main objective of this work is to further advance the research in RHT and to apply it to a variety of applications. The main work can be summarized as the following four parts: 1. Propose the Parameter Space Decomposition Approach to alleviate the high memory requirement. The basic principle of the parameter space decomposition approach is to use several 2-D arrays to implement a high dimension parameter space, which can drastically reduce the space burden. In fact the parameter space decomposition approach can be considered as a trade-off between a large space reduction and a slight false extraction rate. Numerous experiments show that the parameter space decomposition approach is an effective way of RHT implementation. 2. Based on a reasonable criterion, namely the expected number of random samples of minimum subset for a single successful primitive extraction, the performance of the Randomized Hough Transform and the Tabu Search algorithm, the RANSAC algorithm are compared. We show that the Randomized Hough Transform generally outperforms the Tabu Search algorithm and the RANSAC algorithm. In particular, based on a large number of simulations and experiments with real images, we show that with a comparable performance, the Randomized Hough Transform is about twice as fast as the Tabu Search algorithm and the RANSAC algorithm for both line extraction and circle extraction. 3. Proposes a randomized Hough technique to directly extract general triangles (i.e., with unknown size and orientation) from images. Extensive simulations as well as experiments with real images show that the results are satisfactory. 4. Proposes a new Hough Technique for automatic quasar recognition. The experimental results show that the proposed technique is conceptually simple, robust and efficient.
Other Identifier511
Document Type学位论文
Recommended Citation
GB/T 7714
唐珉. Hough变换与几何基元提取[D]. 中国科学院自动化研究所. 中国科学院自动化研究所,1999.
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
[唐珉]'s Articles
Baidu academic
Similar articles in Baidu academic
[唐珉]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[唐珉]'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.