Near Real-Time Volumetric Estimates Using Unmanned Aerial Platforms Equipped with Depth and Tracking Sensors

Abstract

Volume estimation of specific objects via close-range remote sensing is a complex task requiring expensive hardware and/or significant computational burden, often discouraging users potentially interested in the technology. This paper presents an innovative system for cost-effective near real-time volume estimation based on a custom platform equipped with depth and tracking cameras. Its performance has been tested in different application-oriented scenarios and compared against measurements and state-of-the-art photogrammetry. The comparison showed that the developed architecture is able to provide estimates fully comparable with the benchmark, resulting in a quick, reliable and cost-effective solution to the problem of volumetric estimates within the functioning range of the exploited sensors.

Keywords

• close-range remote sensing
• drones
• near real-time three-dimensional scene reconstruction
• volume estimation
• Kalman filtering
• Intel RealSense