基于全球定位系统的相机标定方法研究

A GPS-based Camera Calibration Method

采用标定模板进行相机标定是目前广泛应用的方法。由于标定模板的限制,该方法在大视场、远距离工作时不能覆盖所需画面的全部,不能实现对相机的准确标定。结合全球定位系统实时动态(GPS RTK)定位工作模式的高精度全方位坐标信息获取技术,提出一种基于GPS的相机标定方法。该方法基于GPS RTK工作模式获取GPS质心在WGS-84坐标系中的坐标信息,通过坐标转换得到世界坐标系中的坐标位置;用被标相机拍摄GPS在不同位置的图像,采用Harris角点检测算法获得GPS质心的图像坐标,根据成像模型及相机投影矩阵,实现相机标定。实验证明,该方法可有效获取全方位的空间坐标信息,与传统相机标定方法相比,图像像素标定均方误差仅为0.125 36,是一种行之有效可取代标定模板的相机标定方法。

The calibration templates have been widely used for camera calibration. The calibration templates could not cover a large field of view at a long distance due to its limitation. The method could not realize precise calibration. Combining the high-precision coordinate information access technology with GPS RTK mode, a GPS-based camera calibration method is proposed. In the proposed method, the coordinate information of GPS centroid in WGS-84 coordinates is obtained under the GPS RTK mode. The coordinate system transformation method is used to transform WGS-84 coordinates to system world coordinates. The images of GPS at different positions are taken by the camera under calibration, and Harris corner detection algorithm is used to get the corresponding image coordinates. The projection matrix is calculated according to the camera model, and the camera calibration is realized. Experimental results show that the proposed method can be used to obtain omnidirectional spatial coordinate information with high precision. Compared with traditional camera calibration methods, the calibrated RMS error of the proposed method is only 0. 125 36.