基于主动视觉的摄像机线性自标定方法

摘要

为了解决传统摄像机标定方法的精确度差和使用不方便，本文提出了一种基于主动视觉的摄像机线性自标定方法，该方法通过控制摄像机做一次纯轴向运动来确定主点坐标，再做一次纯旋转运动利用单应性矩阵与旋转矩阵的关系确定摄像机的其他三个内参数。这种基于主动视觉的自标定方法原理简单、实现方便、可操作性强。实验结果表明，该方法能够比较准确地标定出摄像机的所有内参数。摄像机标定是计算机视觉中的一项基本任务。传统的标定方法是通过结构的物体（如标定块，标定板等）在图像中的投影计算摄像机的内参数。但是这种方法必须要有标定物，在很多实际应用中难以实现。为了解决传统摄像机标定方法的精确度差和使用不方便，%To solve the problem of inaccuracy and inconvenience of the traditional camera calibration, a new active vision based linear camera self-calibration approach is proposed in this paper. The coordinate value of the principal point of the camera can be calculated linearly by controlling the camera to undergo a pure axial motion and the intrinsic parameters of the camera can be calculated linearly by controlling the camera to undergo a pure whirl motion. This method owns the characteristics of simplicity in theory, case in implementation and good maneuverability. Experimental results show that the intrinsic parameters can be calculated precisely by applying this technique. Camera calibration is a fundamental computer vision tasks. The traditional calibration method is the structure of objects (such as block calibration, calibration plate, etc.) in the image, the camera intrinsic parameters of projection calculation. However, this method must have a calibration object, in many applications difficult to achieve.