基于相机标定的图像拼接算法  

Image Mosaic Algorithm Based on Camera Calibration

作　　者：张金玲[1] 孙汉旭[1] 贾庆轩[1] 刘亚辉[1] 

机构地区：[1]北京邮电大学自动化学院,北京100876

出　　处：《中北大学学报（自然科学版）》2008年第6期575-579,共5页Journal of North University of China(Natural Science Edition)

基　　金：国家863计划资助项目(2007AA01Z329);国家自然科学基金资助项目(60803103)

摘　　要：针对现有图像拼接算法在配准精度和速度方面的不足,提出了一种利用相机标定信息和相位相关技术相结合的图像配准方法,并给出了算法的详细推导过程.算法通过相机标定获得投影矩阵、相机坐标系间的旋转变换矩阵和平移向量.利用相机坐标系间的旋转矩阵推导对应图像间的旋转变换公式,通过z轴位移及场景的平均景深近似求解图像间的缩放系数,利用相位相关法求解图像间的平移转换参数.实验结果证明该算法理论推导正确,在有效景深范围较小情况下图像拼接准确,对于分辨率为640×480的输入数据,平均拼接速度能够达到26.8 m s.An image mosaic solution based on camera calibration and phase correlation is put forward to improve the accuracy and speed of image matching in stitching algorithm, the deducing steps are introduced. Camera calibration is done to get internal projection matrix of cameras and the rotation transformation matrix and translation vector between two camera coordinate system. Rotation transformation matrix is used to analyze the rotation transformation equation between the corresponding image pair, translation value on z axis and the average scene depth are used together to calculate the approximative scaling parameter of image pair, and then, phase correlation algorithm is used to get the translation parameters. The simulation result proved that the theory analysis is correct, and stitching result is good in the instance that the effective depth scale of scene is relatively small. For input data with resolution of 640×480, the average mosaic speed is 26.8ms.

关 键 词：图像拼接 相机标定 相位相关 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]