相机内参数精密校准技术的研究

发布时间：2018-05-28 13:35

本文选题：非参数模型 + 校准平台　；参考：《中国民航大学》2017年硕士论文

【摘要】：近年来随着航空航天、大型舰船、高速动车等大型装配制造业对数字化装配要求的不断提高,人们对现场测量任务提出了更严格的要求。视觉测量技术因为具备非接触、实时性强、精度高等显著优点,被当做是实现现场、在线测量的最有效的手段之一。相机内参数校准是视觉测量系统中的重要一环,对整个测量系统的诸多性能造成重要影响,已经成为关键性技术问题。目前相机内参数校准多是采用参数化模型解算的方式,存在很多固有的缺陷,如相机内参数之间的相关性影响、优化求解的非线性不稳定性等,难以满足现代装配制造业的高精度、大空间测量的要求。本文采用纯光学方法精确模拟相机的成像过程,通过构造三维空间成像光束和图像点之间的关系,从而建立像面与空间角矢量的对应,真实地反映并重现相机成像的逆过程,突破具体空间的约束,从原理上实现大空间、高精度测量相机的精确校准。通过研究非参数模型的内参数校准方法,搭建了相应的校准平台,实验验证了方法的可行性,并针对校准控制点的图像处理进行了研究。本文将从以下几个方面进行展开:1.研究了相机常规的内参数校准方案,选择两种常规的校准手段进行初步校准,阐述其校准原理并指出不足。另一方面,针对不同像元尺寸的相机进行模拟实验,分析其对参数模型的影响。2.研究了一种非参数模型的内参数校准方法,分析其校准原理和具体的实施方案,并构建校准平台,同时对影响非参数模型内参校准方法的相关误差进行分析。3.对校准控制点进一步剖析。重点研究红外LED灯作为校准点的优越性,并设计实验,一方面从物距、视场角等角度对校准控制点中心精度的提取做了相关的分析,并得出了相应的结论;另一方面,分析校准控制点的质心定位方法,得到相应的结论。4.采用非参数模型的内参校准方法对实验室相机进行校准,得到了相应的内参数模型。设置交比不变性验证实验,验证非参数模型校准方法的可行性及稳定性都优于常规的参数化模型校准方法,能够有效地提高测量不确定度指标。
[Abstract]:In recent years, with the increasing requirements of digital assembly in aerospace, large ships, high-speed motor cars and other large-scale assembly industry, people put forward more stringent requirements for the field measurement task. Visual measurement technology is regarded as one of the most effective means to realize field and on-line measurement because of its advantages of non-contact, high real-time and high precision. The calibration of camera parameters is an important part of the vision measurement system. It has become a key technical problem and has an important impact on the performance of the whole measurement system. At present, the calibration of camera parameters is mostly based on the parameterized model, which has many inherent defects, such as the influence of the correlation between the camera parameters, the nonlinear instability of the optimization solution, and so on. It is difficult to meet the requirement of high precision and large space measurement in modern assembly manufacturing industry. In this paper, the pure optical method is used to accurately simulate the imaging process of the camera. By constructing the relationship between the 3D imaging beam and the image points, the correspondence between the image plane and the space angle vector is established, and the inverse process of the camera imaging is truly reflected and reproduced. The precision calibration of large space and high precision measuring camera is realized in principle by breaking through the restriction of specific space. By studying the internal parameter calibration method of the non-parametric model, the corresponding calibration platform is built, the feasibility of the method is verified by experiments, and the image processing of the calibration control point is studied. This article will be carried out from the following aspects: 1. This paper studies the routine internal parameter calibration scheme of the camera, selects two normal calibration methods to carry on the preliminary calibration, expounds its calibration principle and points out the deficiency. On the other hand, the simulation experiment of camera with different pixel size is carried out to analyze its influence on the parameter model. This paper studies a non-parametric model's internal parameter calibration method, analyzes its calibration principle and concrete implementation scheme, and constructs a calibration platform. At the same time, it analyzes the relative errors that affect the non-parametric model's internal parameter calibration method. Further analysis of calibration control points. The superiority of infrared LED lamp as calibration point is studied emphatically, and the experiment is designed. On the one hand, the accuracy of calibration control point is analyzed from the angle of object distance and field of view angle, and the corresponding conclusions are obtained. The centroid positioning method of calibration control point is analyzed, and the corresponding conclusion. 4. The calibration method of internal parameter of non-parametric model is used to calibrate the laboratory camera, and the corresponding internal parameter model is obtained. The feasibility and stability of the non-parametric model calibration method are verified by the cross-ratio invariance verification experiment, which is superior to the conventional parametric model calibration method, which can effectively improve the uncertainty index of measurement.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG806

【参考文献】