Here is the mission. The area shown in RED needs to be mapped with 1′ contours and rectified orthomosaic. How many Ground Control Points (GCPs) are needed and what would be the the best way to fly the mission?
Choosing the right drone flight pattern

Three different flight patterns were flown to determine which method is better. A, B or C?



Several permanent Ground Control Points (GCPs) and check points were placed along the dam for testing purposes.

Six GCPs were selected for balanced aerial triangulation. Then the camera optimization was ran to calculate the RMSE values.

The site was flown with a Mavic 2 Pro. Rolling shutter compensation was turned off for the initial comparison results.
Below is the result for pattern A.

Below is the result for pattern B.

Below is the result for pattern C.
Note: GCP 1 & 2 was not included in flight pattern C.

Reviewing the Check Points RMSE values, flight pattern C yields the best result for the horizontal position.
Flight pattern B yields the best result for the vertical position.
———-
Reviewing the Control Points RMSE values, flight pattern C yields the best result for the horizontal position.
Flight pattern C yields the best result of the vertical position.
———-
For this particular mission, flight pattern C was used for the final results.
All control points were used as GCPs.
Without rolling shutter compensation.
With rolling shutter compensation.
Final 1′ contours
The flight pattern along with the right number of GCPs placed in strategic locations is the best solution for 3D survey mapping.
There is always going to be floating error and one solution may not work with every flight scenario.
The trick is to look at the mission flight area and plan accordingly to determine the best approach to minimize any errors.
Different sensors will result in different solutions. Not all drones/sensors are equal.
Remember the deliverables are the key factor. Select the sensor that will meet the deliverables. Select the drone platform that can carry the sensor.
Each sensor/lens combination will have distortion. Calculate the Neat area for each sensor to be able to determine the right sensor for the deliverables. Different sensors will need a different number of GCPs in strategic locations to meet the deliverables.
The higher the micron value “µ”, the more information that can be stored per pixel. Anything above a 1.5 µ is recommended for 3D mapping.
NOTE: The higher megapixel does not always mean a higher µ value. See the chart below.
Conclusion? When it comes to choosing drone flight pattern, doing it right takes some work and planning.
###
LearnCST contributor Jim Crume created the Survey Mathematics Made Simple and Survey Mapping Made Simple book series to showcase the technical aspects needed to become a professional design level mapper.
Previous Article: How to Start a Design Level Drone Mapping Business