3D mapping techniques have a large variety of applications from entertainment to military and medical fields. However, there is a big challenge of obtaining well refined 3D model from a set of images without usage of depth sensors. In the paper, we analyze main components of 3D reconstruction pipeline allowing to get detailed models of outdoor objects from drones. In particular, we experiment with algorithms required for structure from motion and point cloud densification. It is demonstrated, that proper local feature extraction, matching and verification directly effect on a final model quality. Analysis of two existing 3D reconstruction frameworks (MVE and COLMAP) is conducted. Initial experimental results are shown.
Outdoor Mapping Framework: from Images to 3D Model