针对全景系统中的大视场红外摄像机和可见光摄像机带来的图像畸变问题,在分析了畸变原理和数学模型的基础上,首先采用模板法标定摄像机的畸变参数,然后对畸变图像中像素点位置进行几何变换,再对像素点上的灰度值进行双线性插值法灰度校正,最后通过优化算法提高了图像运算速度。试验结果表明,设计的畸变校正算法能实时的校正全景系统中的畸变图像,有效提高了全景图像的观察效果。

A radial distortion correction algorithm is proposed for the distorted image caused by wide FOV of cameras in panoramic system. After analyzing the principle and mathematical model of image distortion, the distortion parameters of cameras are calibrated by template method, then the geometric transformation of the pixel positions in the distorted image is carried out, and the bilinear interpolation of pixels gray value is also used to achieve the anti-aliasing capabilities for transformation image. Finally, the speed of image processing is improved by optimizing the algorithm. The experiment results show that the proposed method can correct the distortion image in real time, and the observation effect of panoramic images has been significantly improved.