Fast 3D Reconstruction of Indoor Scenes Using Height-Adjustable Mobile Lidar

Abstract

To enhance the model integrity in the fast three-dimensional (3D) reconstruction of complex indoor scenes, we propose an automatic acquisition method of point cloud data and RGB images at different heights, using height-adjustable mobile LIDAR (a system for light identification detection and ranging) for indoor scenes and design a complete reconstruction workflow. In the data acquisition and processing phases, some strategies and algorithms are used to eliminate the redundant data of a point cloud, which can effectively reduce the computation burden of subsequent processes such as the point cloud registration. Our experimental results further demonstrate that this method not only enhances the integrity of indoor scene models and visual effects of model roaming at different heights but also improves the speed and accuracy of 3D reconstruction.