Georeferencing Hovermap Point Clouds using ‘Best-Fit’ Sphere Registration in Cloudcompare.
Dr. Jeremy J. Sofonia, CEnvP – Solution Manager, Emesent
If you’re not aware, we recently added the incredible Hovermap system by Emesent to our product line up. Hovermap is packed full of smart features to help you get the best data in challenging areas. Designed to be versatile and lightweight, Hovermap is cable of being mounted to drones, vehicles and backpacks for walked surveys.
However to be considered a useful asset in your arsenal, Hovermap needs to be put through it’s paces and comparison drawn with other solutions. This week we’re brining you an Accuracy Evaluation report published by Emesent detailing the capabilities of the Hovermap system.
Understanding performance strengths, and limitations, of any scanning system is critical to achieving project goals regardless of how it is deployed. In many scenarios, this is achieved through a comparison of the point cloud(s) obtained to accurate ‘known’ control points.
The use of laser scanning spheres for this purpose is well established across the industry as they facilitate accurate georeferencing, co-registration of multiple scans and can provide valuable baseline for quality assurance.
Here, we examine the accuracy of Hovermap at a working quarry through the deployment of laser scanning spheres as control points and a ‘best-fit’ registration method using the CloudCompare software package.
Results from this show that the Hovermap data fits the survey control with an RMSE of 14mm and a mean pair-wise 3D distance measure of 13mm ± 4mm from ground truth.
CONTROL AND DATA CAPTURE
Upon arriving to site, four control points were established across the study area by a licensed surveyor and tripod-mounted prisms.To better visualise the control points within the point clouds during Hovermap data collection, each prism was replaced with a 0.150m Ø polystyrene spherical target. These spherical targets were designed with a fixed vertical offset that coincided with the prism position (Figure 1).
Following the establishment of control, three scans were performed using the Emesent Hovermap LiDAR system. This included two separate flights over the target area with Hovermap mounted
to a DJI M210 (‘drone’) and one walking scan within and around
a crushing plant (Figure 2). Point clouds of the three scans were created and merged into a unified model using the Emesent software (Figure 3) and opened in CloudCompare.
DJI Matrice 300 RTK Spotlight – Secrets & Lies
October 14, 2021
How Drone Data is shaping the future of Asset Management: Trendspek
October 13, 2021
Drones in Filmmaking – The best drones for the job
October 07, 2021
LiDAR in Public Safety
October 06, 2021