平行摄影序列影像的核线影像生成方法

A method for epipolar image generation in parallel photogrammetry sequence image

针对平行摄影方式较传统航空摄影方式获取的影像在核线排列及采集方面存在较大差异,导致平行摄影序列影像难以生成核线影像的问题,该文在研究平行摄影方式的摄影基线与影像核点位置间几何关系的基础上,推导了平行摄影序列影像核点位置的严密计算公式,并根据相邻平行摄影序列影像重叠区域最大化原则选取核线生成的起始点和终止点,以此确定核线影像的生成区域,最后采用灰度双线性内插法完成核线影像的采集。以勘探洞平行摄影序列影像作为实验对象,首先基于该文算法生成其核线影像,然后采用密集匹配算法建立左、右核线影像像素间对应关系,最后结合前方交会算法生成洞室的三维点云,进而验证了该文核线影像生成算法的正确性。

Aiming at the problem that compared with traditional aerial photography methods,the parallel photography method has great differences in the arrangement and collection of epipolar line,which makes it difficult to generate epipolar images in parallel photography sequence images.In this paper,the rigorous calculation formula for the position of the epipolar line of the parallel photographic sequence image was deduced,based on the study of the geometric relationship between the photographic baseline and the epipolar line location in parallel photography.The starting point and the ending point of the epipolar image generation were used to determine the generation region of the epipolar image.And the starting point and the ending point was selected based on the principle of maximizing the overlapping area of adjacent parallel photographic sequences.Finally,the epipolar line image acquisition was completed by gray-scale bilinear interpolation.Taking the exploration hole parallel photographic sequence image as the experimental object,firstly,the epipolar image was generated based on this algorithm,then the pixel correspondence between the left and right epipolar image was established by using the dense matching algorithm,and finally the 3 D point cloud of the cavern was generated by combining the forward intersection algorithm,the correctness of the image generation algorithm for epipolar lines in this paper was verified.