基于虚拟靶标的摄像机标定

Camera calibration based on virtual target

针对视觉测量中的摄像机标定问题,提出利用高精度三坐标测量机构建虚拟立体靶标的方法进行摄像机标定,并将Zernike矩亚像素边缘提取方法与改进后的最小二乘拟合算法相结合完成靶点中心定位。实验表明,该方法提高了靶点中心定位速度以及定位精度,所提出的靶点中心定位方法的定位精度可达到亚像素级,标定结果准确。

A camera calibration method using the high-precision Coordinate Measuring Machines to build a virtual target is presented. The target image center location is achieved by applying Zernike moment to extract target image edge and improved least square method to calculate the ellipse center. Experiments show that the speed and accuracy obtained by the proposed method is much higher than the traditional one. The proposed center location method can improve the location accuracy to sub-pixel level with precise calibration result.