基于视频的摄像头实时标定算法  

Camera Real-time Calibration Algorithm Based on Video

作　　者：陈滨[1] 王毅[1] 赵建军[1] Chen Bin Wang Yi Zhao Jianjun(Department of Weapon Science ＆ Technology, Naval Aeronautical Engineering Institute, Yantai 264001, China)

机构地区：[1]海军航空工程学院兵器科学与技术系,山东烟台264001

出　　处：《兵工自动化》2017年第1期42-45,54,共5页Ordnance Industry Automation

摘　　要：针对人为选取标定图像具有较大的不确定性问题,提出一种基于视频的摄像头实时标定算法。介绍摄像机模型,通过初始图像生成模块采集初始标定图像集合,根据误差分析模块的输出进行标定图像更新,结合精度分析模块判定算法是否结束,并进行实验验证。实验结果表明:该方法能够大幅减少标定图像集合的误差均值及方差,有效地得到标定图像集合,提高摄像头参数的精确度。As the images chose artificially may lead to uncertain results, proposed a camera real-time calibration algorithm based on video. After introducing camera model, the initial calibration images set is acquired by initial image generation module. Carry out calibration image renew by error analysis module output, combine accuracy analysis module to judge the algorithm is end or not, then carry out test to verify. The test results show that the method can greatly reduce the error mean and deviation of calibration image set, effectively acquire calibration image set, and improve precision of camera parameter.

关 键 词：实时视频 摄像头 参数标定 误差分析 精度分析 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]