基于矩特征和鲁棒性估计的摄像机标定方法  

A method of camera calibration based on moment features and robust estimation

作　　者：喻夏琼 秦晓东 高超 孙韬 YU Xiaqiong;QIN Xiaodong;GAO Chao;SUN Tao(Troops 32021, Beijing 100094, China;Space System Division, Strategic Support Troops, Beijing 100120, China;Troops 61618, Beijing 100094, China;Key Laboratory of Technology in Geo-spatial Information Processing and Application System, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, China)

机构地区：[1]32021部队,北京100094 [2]战略支援部队航天系统部,北京100120 [3]61618部队,北京100094 [4]中国科学院空天信息创新研究院空间信息处理与应用系统技术重点实验室,北京100190

出　　处：《测绘工程》2021年第6期16-20,26,共6页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(61701477);国家重点研发计划(2017YFB0502903)。

摘　　要：针对当前平面标定方法中特征点提取方法主要为手动或者半自动方式,提出一种基于普通圆形靶标的摄像机快速标定方法。文中利用基于矩特征的椭圆中心检测方法获取包含外点的靶标特征点集;基于特征点邻域的拓扑结构构建矢量集,构建有序的图像匹配点对;利用鲁棒的参数估计方法精确求取单应矩阵。实验表明,该方法可以自动准确提取液晶显示屏上的圆形靶标的投影点,并实现图像靶标点与真实坐标的自动准确对应。利用该方法,可快速准确自动标定出摄像机内外参数和畸变参数,适用于摄像机的现场快速标定。According to the fact that feature points are detected and corresponded in plane calibration mainly in a maual or semi-auto way,a fast field camera calibration method based on common circular target is provided.Ellipse center feature point sets within outlier are detected by a moment features detection method.Vector set is constructed based on neighborhood topological structure of feature points,and ordered image matching points are acquired.Homography matrix are calculated by a robust parameter estimation method.Experiment shows that the method can detect the projection points of circular target in LED screen,and the image target points and the real coordinates are corresponded in an automatic and accurate mode.The internal and external parameters and distortion parameters can be calibrated through this method quickly and accurately.The method is suitable for fast field camera calibration.

关 键 词：摄像机标定 平面靶标 单应矩阵 鲁棒性估计 LM算法 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]