3D Reconstruction of Oil Refinery Buildings Using a Depth Camera

Abstract

We propose a method for fast three-dimensional reconstruction of oil refinery buildings and facilities using photography with the help of an unmanned aerial vehicle (drone) and a depth sensor for depth image classification. We propose two workflows for object reconstruction for different applications. To use the depth sensors, we have developed a method combining the Structure from Motion (SfM) method for determining the structure of an object from imaging of motion in passive vision mode and the Structural Light (LS) method or Time-of-Flight (ToF) method in active (illuminated) vision mode. With low costs, the SfM method combined with a depth sensor let us obtain images large-scale objects at high speed, where the number of calculations is significantly reduced, the efficiency of 3D reconstruction is improved, and the uncertainty in determining the object parameters is not more than a few centimeters.