An Accuracy and Repeatability Analysis on the Approximation of Non-ideal Roof Features by Ideal Planes

The alignment of large datasets associated with 3D reconstructions is still an open issue in the scientific community. Among the several strategies available in the literature, some are based on using geometrical features. In this paper, the Authors have investigated the reliability of associating ideal planes with non-ideal roof features, which may be used to realign 3D terrestrial reconstructions. The obtained results underline a deep influence from noise, which affects the non-ideal roof feature and ideal plane association, and the necessity to define additional experiments to confirm the robustness of the solution.