多摄站普通数码影像两种量测算法的比较

Comparison of Two Algorithms of Multi-station Photogrammetry Based on Non-metric Image

多摄站量测算法具有地面摄影多像片重叠范围大、定位精度及匹配可靠性高等优点,本文采用单像后方交会——多片前方交会及多片光束法两种算法对基于普通数码影像的多摄站量测算法进行了探讨,并对其解算精度及可靠性进行了比较。实验结果表明,经过相机内方位元素标定、影像畸变校正之后,两种算法可行,控制点较少时,摄站数增加对提高精度作用明显,检查点点位中误差可减小约1mm,且3站以上多摄站摄影测量网可视为稳定。多片光束法对初值要求苛刻,收敛性较差,但精度高于多片前方交会。

Multi-station photogrammetry can make full use of the images that might be taken in all directions of an object,thus the solution is more accurate and the automatic multi-baseline stereo matching is more reliable.This paper has studied the ＂space resection——multi-forward intersection＂ and bundle adjustment based on non-metric images,using two algorithms of multi-station photogrammetry.The accuracy and reliability of the two methods were analysed.The result indicates that after calibration and distortion compensation of non-metric images,the two algorithms are practical.The accuracy and reliability of two algorithms are improved as the number of camera station adds,and when there are few control points,the accuracy rise is more obvious,the mean square error of check point can be reduced about 1mm.When there are more than 3 camera stations,the multi-station net is considered to be stable.Compared with the forward intersection,bundle adjustment requires initial value more strictly,and is hard to be convergent,but the solution of it is more accurate than space resection-forward intersection.