通过对整个计算机监控系统的误差源进行详细分析,将这些误差划分为制造误差、测量误差、计算误差三大类。在此基础上,提出了一种高精度的标定方法。此方法将标定过程明确划分为标定数据的提取、标定模型的建立、摄像机内外参数的计算和优化、畸变校正四个阶段。随着标定过程的推进,先后对图像像素量化噪声误差、图像坐标轴正交误差、不对心误差、镜头的径向畸变误差和切向畸变误差进行了修正。实验结果表明,标定方法非常精确,最后的标定总误差被控制在0.06个像素单位以内。
This paper analyzed the error sources of computer monitor system. All of them could be divided into 3 sorts : manufacture error, measurement error and computation error. Then, presented a precise calibration method. Calibration procedure was described by 4 steps: extracting calibration data, fitting camera model, computing and optimizing both camera intrinsic and extrinsic parameters, correcting distortion. Quantization noise of pixels, orthogonal error of image axes, asymmetric error, radial distortion and tanlggential distortion of lens were corrected along with executing calibrition. Experiment results Show that this method is precise. Sum of all errors can be controlled within 0. 05 pixels.