近景大视场角分区域标定方法研究

Research on Calibration Method of Sub-area of Large Field of View

针对飞行试验测量视场大相机标定精度低的问题,提出一种高精度CCD相机分区域标定方法;该方法首先通过将标靶均匀布置在摄像机视场内,使得标靶尽可能均匀错落地充满整个视场范围,再结合人眼判读的方式求解靶标的像面位置,最终与全站仪三维坐标形成精确的空间标定点集;接着,将像平面按横向方向等间距分割成N个区域,并结合后方交会的方法分别对每个子区域进行相机参数的计算;实验结果表明:经过分区域标定,相机采集点的总误差比单区域标定法降低了4%(N=3);算法可实现指定区域的相机参数计算,基本满足中高等精度的工业测量要求;所本文研究可应用于位置相对固定不变的工业视觉测量,特别是大工件测量领域。

Aiming at the problem of low calibration accuracy for large -- scale camera in flight test, a high -- precision CCD camera sub -- area calibration method is proposed. The method first places the target evenly within the camera field of view, so that the target as wide as possible to fill the entire field of view. And then combined with the way the human eye to solve the target image plane position. Finally with the total station three -- dimensional coordinates to form a precise spatial calibration point set. Then, the image plane is divided into N regions by equidistant direction in the transverse direction, and the camera parameters are calculated for each sub--region in combination with the method of the intersection. The results of calibration experiment show that the total error of the camera acquisition point is 4% lower than that of the single region calibration method after the regional calibration. Algorithm can achieve the specified area of the camera parameter calculation, the basic to meet the high precision industrial meas- urement requirements. The proposed sub--regional camera calibration method can be applied to the relatively constant position of the industrial visual measurement, especially in the field of lane workpiece measurement.