基于RTCam摄像机模型的桶形失真校正研究

发布时间：2018-01-08 23:33

本文关键词：基于RTCam摄像机模型的桶形失真校正研究　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：空间视觉系统是宇宙空间探测的关键设备之一,对于系统中的每一台摄像机,其在成像过程中都需要建立自己的空间模型对空间图像进行采集,多个摄像机联合探索完成任务时,各自独立的空间模型必须拼接成为完整的经过修正的统一的空问模型。针对不同摄像机的成像模型需要统一建模的问题,英围科学家提出了Rational-Tensor摄像机模型(简称为RTCam摄像机模型),建立了不同类型的非线性摄像机模型的统一约束关系以及几何框架。本文将RTCam摄像机模型应用于图像桶形失真的校正,利用基于RTCam摄像机模型的桶形失真校正算法,对存在严重桶形失真的图像进行校正,并对校正精度进行了定量分析。首先,对普通镜头采集的无失真图像进行RTCam变换,得到无失真图像的存在严重桶形失真的图像,对桶形失真图像计算校正参数S因子的值,利用基于RTCam摄像机模型的桶形失真校正算法对桶形失真图像进行校正,通过对失真前和校正后的特定直线进行拟合,对比拟合的误差参数说明本算法可以将发生严重桶形失真现象的曲线校正成为直线,并对比直线拟合的一次项系数和常数项的值说明校正后该直线与失真前相比在图像平面没有发生偏移,从而证明了算法可以高精度的校正存在严重桶形失真的图像；然后,利用基于RTCam摄像机模型的桶形失真校正算法对广角镜头采集的存在严重桶形失真的图像序列进行校正,对校正后的图像中的特定直线进行拟合,对比拟合的误差参数说明本算法的直线拟合误差低于传统的基于多项式模型的桶形失真校正算法,从而定量的说明了与传统的桶形失真校正算法相比,本算法具有更高的校正精度。
[Abstract]:Space vision system is one of the key equipment for space exploration. For every camera in the system, it is necessary to build its own space model to collect the space image in the imaging process. When multiple cameras jointly explore and complete the task, each independent space model must be stitched into a complete revised unified spatial model. The problem of unified modeling is needed for different camera imaging models. The Rational-Tensor camera model (referred to as the RTCam camera model) was proposed by Yingwei scientists. The unified constraint relation and geometric framework of different nonlinear camera models are established. In this paper, the RTCam camera model is applied to the correction of image barrel distortion. The barrel distortion correction algorithm based on RTCam camera model is used to correct the image with serious barrel distortion, and the correction accuracy is analyzed quantitatively. The RTCam transform is carried out on the undistorted image collected by the common lens, and the image with serious barrel distortion is obtained. The correction parameter S factor is calculated for the barrel-distortion image. The barrel distortion correction algorithm based on RTCam camera model is used to correct the barrel distortion image, and the special straight lines before and after correction are fitted. The comparison of fitting error parameters shows that this algorithm can correct the curve with serious barrel distortion into a straight line. By comparing the coefficient of the first term and the constant term of the straight line fitting, the correction shows that the line has no deviation in the image plane compared with that before the distortion. It is proved that the algorithm can correct the image with high accuracy. Then, the bucket distortion correction algorithm based on RTCam camera model is used to correct the image sequence which has serious bucket distortion, and the special straight line in the corrected image is fitted. The comparison of fitting error parameters shows that the linear fitting error of this algorithm is lower than that of the traditional bucket distortion correction algorithm based on polynomial model, which shows the quantitative comparison with the traditional bucket distortion correction algorithm. This algorithm has higher correction accuracy.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN948.41;TP391.41

【相似文献】