基于彩色同心圆的投影仪标定算法

本文研究主动视觉系统中投影仪的平面标定方法。用固定的投影仪向画有青色同心圆的标定板上投影黄色与洋红色相间的棋盘格,用固定的摄像机采集若干幅标定板不同朝向的图像,对采集图像进行颜色通道过滤,得到投影仪标定所需的同心圆分量和棋盘格分量,结合摄像机投影仪图像之间的映射关系和同心圆的秩2约束来标定投影仪。此方法仅需用三幅图像,不仅可以标定出投影仪的所有内参数(包括倾斜参数),如果再一步深入,还可以标定出摄像机的内参数矩阵和主动视觉系统的外参数矩阵。实验结果表明本文的标定算法精度比较可观。

A colored concentric circles based projector calibration method

We present a planar based method of calibrating a pinhole model projector in this paper. A yellow-carmine checker pattern is projected on a calibration board, on which two cyan concentric circles were drawn, to integrate checker component and concentric circles component into one plane. Change the orientation and position of the calibration board several times to gather a few images of the plane. Separate those two components using color channel technology. Then calibrate the projector using the rank-2 constrain provided by concentric circle component and homography between projected pattern and gathered image provided by checker component. The internal parameters and the external parameters of the active vision system as long as the internal parameters of the camera, if needed, can be acquired by processing only three gathered images. It is a practical calibration method with considerable accuracy.