基于多相机系统标定的全景拼接算法研究

Research On panoramic stitching algorithm based on multi-camera system calibration

全景拼接技术是当前全景技术的重点,而多相机系统被越来越多的应用于全景技术中。文章提出了一种基于多相机系统标定的全景拼接方法,利用通过标定得到的相机系统相关参数,将多相机系统拍摄得到的图像进行畸变矫正与旋转变换处理,以提高全景拼接的效果。使用基于特征描述符的标定板对多相机系统进行标定,获得相机的内部参数和外部参数;利用Mercator投影,结合标定得到的内部参数对鱼眼图像进行畸变矫正,外部参数则用于对矫正后的图像进行旋转变换处理,使它们在同一世界坐标系下;用Harris算法对相邻待拼接图像的特征点进行检测匹配,进而拼接成全景图。通过实验验证,该全景拼接算法具有不错的拼接精度,没有出现误匹配的情况。

Panorama mosaic technology is the focus of the current panoramic technology, and multi-camera system is more and more used in panoramic technology. In this paper, a method of panoramic stitching based on multi-camera system calibration is proposed. By using the parameters of the camera system calibrated, the images taken by multi-camera system are processed by distortion correction and rotation transformation to improve the effect of panoramic stitching. The multi-camera system is calibrated using the character-descriptor-based calibration plate to obtain the internal and external parameters of the camera. The Mercator projection is used to correct the fisheye image with the calibrated internal parameters. The external parameters are used to correct The images are rotated and transformed so that they are in the same world coordinate system; Harris algorithm is used to detect and match the feature points of the adjacent images to be spliced, and then spliced into a panoramic image. Through experimental verification, the panoramic mosaic algorithm has a good mosaic accuracy, there is no mismatch situation.