Higher accuracy in aerial imagery? Topodrone and Matrix Geo collaborate for an...

Higher accuracy in aerial imagery? Topodrone and Matrix Geo collaborate for an answer

SHARE
Do you know what is the main challenge when drone photogrammetry has to compete with accuracy of LiDAR solutions?
 
In drone based photogrammetry solution, you actually need more number of Ground Control Points (GCP) to achieve consistent quality throughout your project area, which means for achieving higher accuracy you  will have to establish GCP’s at every 200-500 M depending on target accuracy level. Putting Additional GCP’s become a time taking, risky and costly activity and increases your project cost.
 
To combat the issues, Topodrone and Matrix Geo have collaboratively introduced a professional aerial survey solution based on DJI Drone+RTK/PPK.

Methodology

drone photogrammetry

STEP 1

Select Test Area

A test area has been selected, outside of No Flying zone of airports (pic.1). The area was about 110 000 sq. m. (11 hectares), covered by only one battery for every flight altitude ranging from 60/100/150 meters.

drone
Area of Interest

STEP 2

Set up check points & base station

photogrammetry
Base station

The base station REACH RS+ GNSS receiver has been installed on the point with known coordinates.

Checkpoints

rover
Measuring coordinates of checkpoints

The Check Points have been placed and measured in RTK mode by second REACH RS+ (Rover) in evenly distributed manner throughout the test area to establish the check control points.

Checkpoints location

STEP 3

Flight Mission & Aerial Image Acquisition

Three missions were planned at 60, 100, 150 meters’ altitude with 85% along and 70 % across overlapping, flights were performed with up to 9 m/s speed.

Flight planning

STEP 4

Data Processing

  • After all the flights, set of raw images and raw GNSS data were downloaded from the drone and the base station.
  • The data was converted into Rinex format and post processed by RTKLIB application and as a result precise coordinates of images exposure were received that were further embedded to photos by TOPOSETTER application.
  • Datasets from different flight altitudes, i.e. 60, 100, 150 respectively were photogrammetrically processed by Photoscan software without using any Ground Control Points.
  • After processing, the checkpoints coordinates were imported into the Photoscan project and checked for the accuracy of each flight.
  • The data has also been checked into Traditional Stereo Photogrammetry workstation and the result was found well within 3-4 CM accuracy.
  • In all the Missions, the results were well within 3-4 CM Accuracy.

Tables for the result against the GCP’s

Conclusion

The above results from flights of different altitudes, show that the accuracies achieved are well within two pixel of GSD and hence passes the criteria of the National Standard for Spatial Data Accuracy (NSSDA). The outcomes so achieved have taken drone photogrammetry to new heights.

Also Read

An Aerial View of the Future – Drones in Construction