逆向工程中基于CCD的曲面测量方法研究

【作者】 吴剑锋；

【导师】 王文； 陈子辰；

【作者基本信息】 浙江大学 ， 机械制造及其自动化， 2005， 硕士

【摘要】 逆向工程作为一种产品创新设计的重要手段,在敏捷制造、虚拟制造中发挥巨人作用。如何快速、准确地获取被测物体表面大量的测量数据,是目前逆向工程研究的重点内容之一。其中多传感器集成的智能化测量系统是未来自由曲面测量系统的主要发展趋势。本文主要研究基于CCD的自由曲面视觉测量技术,并探讨如何利用CCD测量结果(被测物体表面的曲面特征)智能指导高精度测头进行测量,达到高精度、高速度测量要求。 本论文主要内容有: 第一章综述逆向工程的重要作用,简单介绍逆向工程中的现代接触式测量方法,以及基于视觉技术的非接触式自由曲面测量方法,并总结出三维测量系统的发展趋势,最后给出了论文选题的依据。 第二章结合辅助网格技术,提出基于CCD双目立体视觉的曲面测量方法,并阐述该方法的基本测量原理和几个关键技术。 第三章简单介绍常用的CCD标定方法和相应的成像模型,并结合本测量方法的特点,提出一种简便的CCD标定算法。 第四章给出了图像预处理算法,便于准确、快速地得到网格交点的单像素坐标。该算法主要包括图像分割、网格线膨胀腐蚀以及图像细化等部分。 第五章主要研究基于“两步法”的网格节点提取技术和基于网格点相互位置的匹配算法。 第六章在对传统测量方法进行分析的基础上,提出视觉信息指导下基于矩形细分的快速测量方法,并对该方法进行仿真分析。 第七章主要针对论文中提出的测量方法进行实验研究,分析影响测量精度的多个因素,并提出改进措施。 第八章对全文进行总结,并对后续研究作了展望。更多还原

【Abstract】 Reverse engineering plays an increasingly important role in creative product design and shortening product development. At present, how to acquire a large quantity of 3D coordinate points efficiently and accurately has become a critical issue in reverse engineering. It can be predicted that in the near future, the intelligent measuring system with multi-sensors integration technology will be the most promising trend for free-form surface measurement. This dissertation has studied how to utilize CCD in surface measurement and as well, discussed how to apply the results that feature the surface of the object to instruct higher resolution sensors for further measuring.Main points of this dissertation are provided as follow.In Chapter 1, the importance of Reverse Engineering was summarized, the modern surface-measuring methods, such as contact measure method and the non-contact measure method based on vision technology, were simply introduced and the current of the 3D measuring system was summed up, at last we gave the reason that we studied the question included in this dissertation.In Chapter 2, the surface-measuring method based on the binocular stereo system was put forward to meet the practical requirement of this research, then the measuring principle was elaborated and some key techniques were brought forward as well.In Chapter 3, General ways for CCD calibration and the relative camera models were briefly introduced. Then, a simple and effective method for CCD calibration was put forward.In Chapter 4, In order to obtain the 3D coordinates of the object, we must get the "position" of the corresponding grid nodes quickly and exactly. To achieve this goal, the reasonable image processing algorithm was introduced, and the image processing techniques mainly included the image segmentation algorithm, the image erosion algorithm, the image dilation algorithm and the image thinning algorithm.In Chapter 5, the grid nodes extraction algorithm based on "two steps" and the nodes corresponding algorithm based on the relative positions of the grid nodes were introduced.In Chapter 6, after the analysis of the conventional measure ways with contact probes and laser probes, a fast measure method with the rectangle-division technique was brought out under the direction of the "visual information". The validity of this method has been testified through simulation experimentations.In Chapter 7, some experiments about the measuring method were studied and the factors for its error were analyzed. Then a few ways to improve the measure precision were put forward.In Chapter 8, the conclusion is made and the prospect of the future improvement is also given.更多还原