Gross Error Considered Stereo Images Relative Orientation

Abstract

Relative orientation is one of the most important and classic tasks in photogrammetry. Gross error of the corresponding points may be caused by computer automatic image matching, the commonly used method is unsuitable to relative orientation. In this paper, a novel approach of stereo images relative orientation considering gross error was proposed. Firstly the initial elements of relative orientation are computed by direct relative orientation, then the refined elements of relative orientation are obtained by Huber and iterative algorithm in the continuous relative orientation, which can eliminate the negative influence of gross error. Experimental results show that the proposed approach can achieve more excellent performances in accuracy than the commonly used method in the corresponding points with gross error.