Accuracy of Drone Image Based Volumetric Surveys

Abstract

The use of Unmanned Aerial Vehicles (UAVs) in stockpile calculations is becoming more common due to the time savings associated with its use. In this paper, the accuracy of drone-based volumetric surveys using nadir and oblique images were compared with traditional survey methods. Two cardboard pyramids, measuring 14 cubic yards and 5 cubic yards respectively, were used to simulate a stockpile. The pyramids were surveyed with waypoint-assisted flights at 40ft, 125ft, and 350ft altitudes. Additional oblique images were captured at both sides of the targets and field to supplement the nadir flights. The volumes of the resulting models were then computed with ContextCapture software. This study found that adding oblique images to the nadir surveys did not significantly increase the accuracy of the model volumes. Oblique images on their own also yielded similar results to the nadir images. The accuracy of the models at 125ft ranged from 0.35% to 3.03% from the measured values, while the oblique images were off by 0.44% to 8.23% of the actual values.