Relative orientation based on multi-features

In digital photogrammetry, corresponding points have been widely used as the basic source of information to determine the relative orientation parameters among adjacent images. Sometimes, though, the conventional relative orientation process cannot be precisely implemented due to the accumulation of random errors or in the case of inadequate corresponding points. A new relative orientation approach with multiple types of corresponding features, including points, straight lines, and circular curves, is proposed in this paper. The origin of the model coordinate system is set at the projection center of the first image of a strip, and all of the exterior orientation parameters, except φφ and ω of the first image, are set at zero. The basic models of relative orientation with corresponding points, straight lines, and circular curves are discussed, and the general form of a least squares adjustment model for relative orientation based on multi-features is established. Our experimental results show that the proposed approach is feasible and can achieve more reliable relative orientation results than the conventional approach based on corresponding points only.

► We propose a new relative orientation method with multiple types of features.
► Corresponding points, lines, and circles are used to establish the mathematic model.
► The approach is advantageous in robustness, reliability, and availability.
► Strip bending problem caused by accumulation of random error is effectively reduced.