Three-point-based solution for automated motion parameter estimation of a multi-camera indoor mapping system with planar motion constraint

Abstract

Accurate indoor 3D models have become a key prerequisite for various applications. Through state-of-the-art image processing techniques, 3D models can be generated from high quality images captured by off-the-shelf digital cameras. To acquire redundant data and produce real scale models, a multi-camera system can be used. However, dedicated approaches for image-based 3D reconstruction using mapping platforms equipped with multiple cameras have not been fully addressed. Assuming the availability of prior information regarding the platform trajectory, this paper presents a new approach for reliable estimation of system motion parameters between different data acquisition epochs of a multi-camera system. This approach, which assumes planar motion of the utilized platform, provides a three-point closed-form solution. The derived solutions are then incorporated within a modified RANSAC framework for outlier detection/removal. It is worth noting that, different from the existing General Camera Model (GCM)-based solutions, the proposed approach is based on a modified co-planarity model, which is essentially a direct extension of the classic stereo-based relative orientation. Moreover, since the proposed approach only provides a maximum number of four possible solutions for system motion parameters over different epochs, it has better computational efficiency when compared to other existing algorithms. Experimental results from real datasets acquired with different configurations have demonstrated the reliability of the proposed approach in motion parameter estimation for indoor multi-camera mapping systems.