矩形模板下摄像机标定和目标定位方法研究

Research on Camera Calibration and Object Location Based on Planar Pattern of a Rectangle

提出一种利用矩形进行摄像机标定和目标定位的新方法.该方法根据透视投影几何关系中旋转矩阵的单位正交性,由矩形的四个顶点的像构造无穷远平面上绝对二次曲线的像ω的约束方程标定摄像机内参数;目标定位转化为求解摄像机外参数,无需计算图像和空间平面间的单应矩阵,直接由矩形顶点确定.本文方法完全摆脱了复杂的图像匹配问题,所有计算方法完全线性化,标定过程简便.模拟和真实图像实验结果证明了方法的可行性,表明该方法求解精度高、鲁棒性强.

This paper proposes a new approach for computing camera intrinsic parameters as well as the locations of objects. According to projective geometric and the orthogonality of rotation matrix, linear constrains on camera intrinsic parameters are set up via the absolute conic image fo and the images of rectangle vertexes. This paper turns object location into the recovery of camera extrinsic parameters, which can be computed directly from rectangle vertexes, instead of the homography between the image plane and the space plane. The main advantage of this new approach lies in avoiding the correspondences between images, and the algorithm is linear. The new approach is simple and efficient. Extensive tests with real images and synthetic images validate the proposed approach and show that it is accurate and robust to noise.