IOE OpenIR  > 光电技术研究所博硕士论文
Alternative TitleStudy on Calibration and Matching Methods of Digital Structured Light System Suited for Reverse Engineering
Thesis Advisor袁家虎
Degree Grantor中国科学院光电技术研究所
Place of Conferral光电技术研究所
Degree Discipline测试计量技术及仪器
Keyword标定 匹配 数字结构光系统 三维轮廓测量 数字近景摄影测量 计算机视觉 几何标定 光束平差法 光照标定 摄像机标定 投影机标定 点模式匹配 立体匹配 三维点匹配 逆向工程
Abstract逆向工程是一种新颖的几何造型理念,其技术实现依赖于对产品样件的三维轮廓的精密测量。为了获得任意复杂外形的物体曲面轮廓,各种结构光系统成为了逆向工程中的主流测量设备,以满足逆向工程无损、快速、精确测量的要求。 当前,基于数字光投影技术的数字结构光系统由于其优异的性能,受到了广泛研究。 由同一光线上的三维物点和二维像点组成的点集称为同名点对。按照数字近景摄影测量理论,同名点对是数字结构光系统处理的基本数据。系统的测量过程包含了匹配、标定和求解三个步骤。标定和匹配过程紧密联系,必须迭代进行。 本文使用了基于液晶投影技术的数字结构光系统,研究了标定和匹配中的问题,改进并提出了一些新方法,主要包括: (1) 提出了一套基于数字近景摄影测量原理几何标定数字结构光系统的方案,能够获得高精度的系统几何参数向量,达到了1/12000的相对标定精度。 (2) 建立了数字结构光系统的光照度模型,提出了一种摄像机和投影机的光照标定方法,提高了相移法的相位测量精度。 (3) 针对在低投影光照时双周期相移法的解包裹问题,提出了一种混合解包裹算法,同时使用了空间域和时间域的约束条件。 (4) 实现了一套自动检测和匹配人工标志物的方案,这包括:改进了基于正交接触点算法的椭圆几何拟合方法,采用正交方向上的灰度梯度加权待拟合的边界点数据集,提高了椭圆中心的拟合精度;基于Hough变换来检测位域圆环编码标志物的方法,能够减少噪声块的影响;基于射影不变量来匹配标定板上的平面标志物,确保了无误匹配;给出了一种解决立体视觉匹配和三维点匹配问题的约束优化方法。 最后,测得了样品零件的点云数据,重建了表面轮廓,分析了测量精度,验证了上述方法的有效性。
Other AbstractReverse engineering is a novel idea for geomtric modeling, based on the high precision measurement of sample products. To meet with the requiring of contactless, fast and accurate measuring, various structured light system become majority of the measuring devices, which can acquire surface model of any complicated profile of objects. Currently, digital structured light system (DSLS) is widely researched for outstanding performance, in which digital lighting projection technique is used. Homologous points (HPs) are point set consists of 3D object point and 2D image point which are on the same ray. According to principle of digital close-range photogrammetry (DCRP), HPs are basic data for DSLS. The whole data processing of DSLS consists of matching, calibration and caculation. The procedures of matching and calibration, which are tightly related, must be iteratively executed. On the basis of the excogitated DSLS using TFT-LCD projection technique, some main issues of calibration and matching are studied in this paper. Several methods are novelly improved and proposed, which follows: (1) An approach to geometric calibration of the DSLS is presented, which is completely based on DCRP theory. The high accurate geometric parameters are computed through the approach, where the relative accuracy can be achieved 1/12000. (2) The irradiance model of the DSLS is establised. Base on the model, a photometric calibration scheme of camera and projector is proposed, which improves the accaracy of phase measuring in the phase-shifting method. (3) In the case of low projective irradiance, a hybrid phase-unwrapping algorithm is proposed, where the spatial and temporal constraints are simultaneously used. (4) A frame of auto-detecting and auto-matching manmade feature object are implemented, including: The cost function in the orthogonal contact point algorithm is weighted by the orientation gradient of gray value, which can improve the fitting accuracy of ellipse center. The Hough transform is used for detecing bitset-cirque coded feature that decreases the influence of noisy image blob. To ensure no wrong matchings, the features on the calibration board are matched with the invaraint of perpective projection. A method of resolving the problems of the stereo matching and 3D point matching is introduced, which is using dynamic programming. Finally, the point cloud data of sample part are measured, and the surface model is reconstructed. The precision of result is analysed, which shows the effiecency of proposed methods.
Document Type学位论文
Recommended Citation
GB/T 7714
罗剑. 适于逆向工程的数字结构光系统的标定和匹配方法研究[D]. 光电技术研究所. 中国科学院光电技术研究所,2007.
Files in This Item:
File Name/Size DocType Version Access License
10001_20041801512101(4939KB) 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[罗剑]'s Articles
Baidu academic
Similar articles in Baidu academic
[罗剑]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[罗剑]'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.