Towards Real-Time 3D Monocular-Based Reconstruction in Support of Uas Missions

This paper discusses the processes and resources needed to develop three-dimensional (3D) reconstructions of an object or indoor/outdoor environment using unmanned vehicles equipped with a single camera. A single camera captures a scene in twodimensions (2D), inherently providing no information about the 3D space of the scene. Converting a stream of 2D images into a high-resolution 3D model requires a sequence of coordinated algorithmic processes. Literature review reveals the attempts to tackle the problem by configuring unmanned vehicles with a suite of sensors. This approach solves some problems with scene capturing but opens a large envelope of vulnerabilities when using the system in critical missions requiring quick response, high confidence level and precision.