Validating AeroPoints accuracy
In August Propeller Aero introduced AeroPoints, the world’s first smart ground control points. Fred Greer gets to the point. Since then we’ve had a lot of interest in how the accuracy of AeroPoints compares with conventional methods of establishing ground control for drone surveying.
We’ve had a lot of interest in how the accuracy of AeroPoints compares with conventional methods of establishing ground control for drone surveying.
Traditionally, obtaining ground control data was a costly and time consuming process. A surveyor would mark points visible from the air around the site and measure their position with an RTK rover, an expensive dual-frequency GNSS receiver with a cellular connection back to a network of reference stations at known locations. By comparing its GNSS signal to that from the reference station, the RTK rover can obtain positional data accurate to a few centimeters in close to real time.
The secret ingredient that makes RTK rovers so expensive is the dual-frequency receiver. By swapping this for a single-band receiver, we can make our AeroPoints affordable enough to sell as a set.
There is some folklore around the drone world that single frequency GNSS receivers can only produce accurate positions less than 10km from the reference station when using real-time or post processing algorithms.
We set out to determine exactly how far from a base station you can be while still obtaining centimeter-grade data.
Experimental method
To determine the accuracy of AeroPoints we compared them with a TopCon HiPer V RTK Rover, an industry standard survey-grade multi-constellation, multi-band GNSS receiver that claims measurement accuracy better than 3cm. We took this to be the “gold standard”.
We performed a routine drone survey. We laid out 12 AeroPoints, activated them and measured the center of each with the TopCon unit. We then flew the site to capture our photogrammetry data. After the flight we collected our AeroPoints and uploaded their observation data for automatic post-processing on the Propeller platform. Overall, during this survey, we had captured about 45 minutes of AeroPoints data.
Results
We post-processed the AeroPoints data from our survey against the 13 nearest base stations, which ranged from 7 to 44 km away. We plotted the root mean square (RMS) deviation of our ground control points from the TopCon measurement against this baseline distance.
The results are as follows:
Analysis
There is a slight increase in deviation as the baseline is extended, but the data still meets the definition of “survey-grade” (3 cm X/Y, 5cm Z) out to a baseline distance of over 40 km.
For more information about AeroPoints, visit propelleraero.com/aeropointsor drop us a line at hello@propelleraero.com
Glossary
Phew, that was a pretty technical post! Here are a few definitions:
Baseline distance: The straight-line distance between the reference station and rover receivers in a differential GNSS configuration
GNSS: Global navigation satellite system, The catch-all term for geographic positioning satellite systems like GPS, GLONASS, BeiDou and Galileo
Ground control: The process of gathering a set of known geographic points on a terrain to calibrate the dimensions of a photogrammetric model
RMS: Root mean square, a statistical measure of the average error in a set of data
RTK: Real time kinematic, an algorithm for accurately calculating positions in real time given GNSS observation data from a rover and a reference station at a known location
