An autonomous indoor exploration robot rover and 3D modeling with photogrammetry

Abstract

© 2018 IEEE. An autonomous robot rover has been designed and developed. The proposed robot rover is capable of self-localization and self-exploration of a room or an area inside a building. Self-driving and self-navigating in unknown environments, our robot constantly scans its surrounding, incrementally and dynamically constructing a global map of the visited room. Immediately after it detects an obstacle or an obstruction, the robot car does not only stop or diverge its route in order to avoid crashing into it, but also marks its location on the map as dangerously hazard. All unoccupied spaces are registered as a safe path. While exploring, the robot is programmed to take a large number of photographs of its surrounding, which includes multiple views of each scene from multiple different angles. These images collected are used to create a three-dimensional model of the interior of the room by photogrammetric methods. In addition to a general exploration of desired regions, a robot may be presented with a selected target object. It must then chooses its own path to locate the object and to circle around it in order to take many photographs of it from all directions and to create a corresponding 3D model.