论文部分内容阅读
本文给出了一种基于圆曲线的摄像机自定标的新方法。设计了一个含有三个两两不包含的圆(其中圆心和半径大小是未知的)的平面模板,让摄像机在三个以上的不同方位观察该模板,然后利用模板的图像来求解摄像机的内参数。文中,首先介绍了三维摄影空间中绝对二次曲线和圆环点的相关知识,接着推导出摄像机内参数的约束方程,最后又详细讨论了圆环点及其图像的求法。其中关键技术是通过求解像曲线的切线确定圆心图像的像素坐标,然后确定圆环点图像坐标来实现摄像机的定标。模拟实验和真实图像实验也表明,本方法具有很好的鲁律性。由于所用模板中的圆不需要精确定位,因此该模板易于制作。由于该技术对摄像机的运动没有限制,因此,它可以在一般的实验平台上实现。
This paper presents a new method of self-calibration of camera based on circular curve. A plane template with three non-inclusive circles (where the size of the circle center and the radius is unknown) is designed so that the camera can observe the template in more than three different directions and then use the template image to solve the camera’s internal parameters . In this paper, the knowledge of absolute conic and circle point in three-dimensional photographic space is firstly introduced, and then the constraint equation of the parameters in the camera is deduced. Finally, the method of calculating the circle point and its image is discussed in detail. One of the key technologies is to solve the camera calibration by finding the tangent of the image curve to determine the pixel coordinates of the image of the center of the circle and then determining the image coordinates of the circle point. Simulation experiments and real image experiments also show that this method has good robustness. The template is easy to make because the circle in the template used does not need to be precisely positioned. Since this technology has no limitations on camera motion, it can be implemented on a common experimental platform.