基于平面网格模型的摄像机镜头畸变校正技术

Technology of Proofreading the Camera Lens Distortion Based on the Plane Gridding Model

摄像机标定技术是计算机视觉研究中的关键技术之一。国内外学者开展了大量的研究并取得了大量的研究成果。目前,摄像机标定技术的主要研究任务在于根据实际应用的特点,寻找简便、快捷、准确的标定算法。该文针对精密视觉测量这一特殊应用领域,提出了基于平面网格模型的摄像机镜头畸变校正技术。除了考虑镜头的径向畸变之外,还包括离心畸变、薄凌镜畸变等非线性因素,在镜头中心附近小区域零畸变的假设条件下,推导出网格交点的理想坐标,并由图像处理技术获得网格交点的实际坐标,通过最小二乘法获取非线性畸变系数。由于考虑了引起镜头畸变的多种因素,可提高实际标定的精度。对于视觉测量系统,在具体标定前先求出畸变系数,再对实测的像素进行校正,这样既可以提高标定精度,又不影响视觉测量的速度。

Calibration of camera is one of the key technologies of Computer Vision.Many academicians have done much work on it and acquired a lot of fruits.Now,mostly research is to find a simple,timeless and exacted arithmetic for practical application.Taking into account of exacted vision measurement ,this paper brings forward a method which bases on plane gridding model to proofread the aberrant cameral lens.Besides radial distortion,there are other non -linear factors,such as centrifugal distortion and thin-prism distortion and so on,on the assumption that there is no distortion in the little area around center of the lens,it can infer the actual coordinate of point of the gridding intersection through image-processing technology,then get the non-linear distortion coefficient through least-squares method.For the system of vision measurement ,getting the distortion coefficients before calibrating and measuring,then proofreading the pixel by these coefficients,thus can improve the precision of calibration without affecting the speed of vision measurement.