VOLUMETRIC ESTIMATION OF STOCKPILES USING UAVs

Abstract

Volume estimation is a very important part of many industries and businesses; therefore, it becomes vital to do so quickly and accurately, while ensuring cost efficiency as well. Inaccurate estimation results might hamper businesses and their relationships. Conventional surveying method using tapes, Total station, GNSS and in some instances where accuracy is of utmost importance Laser Scanners, have been widely used for volume computation in surveying for a long time, irrespective of its lengthy, risky and laborious data collection on sites. UAV platforms have the potential to bridge the gap between traditional techniques by providing volume estimates quickly, in a cost effective manner. Although some of the most widely used UAV photogrammetry software are commercial and therefore mostly limited to some prestigious laboratories or big industries only. This project explores the use of Open Source software packages in volume estimation through UAV photogrammetric techniques. A stone crusher site in Laksar was chosen for the task, in which 7 stockpiles were selected for volume estimation. Volume estimation of stockpiles was first performed wholly on open source platforms which included WebODM and CloudCompare, where it explored the use of SfM (Structure from Motion) and CSF (Cloth Simulation Filter) algorithms in obtaining the results. Next the volume computation was performed on one of the most widely used commercial photogrammetry software Pix4D (both Mapper and Survey) to obtain results. The results obtained through both these methods were compared and validated against the ground truth data. The accuracies were assessed and the results were impressive, bringing to a conclusion that there are good alternatives available to commercial photogrammetry software packages in the volume estimation domain. Future scope of work lies in making the results more accurate and closer to the ground truth values as well as working on the automation of computation results.