Illogical GSD vs overlap results

[ArnauCM]

Hi, fellow metashapers

A collegue of mine did some flights for his master thesis and ended up with some apparently illogical GSD results. I have tried to find some logic into it but couldn't.

He combined different tilt angles with overlap percentatges. The results are shown in the table image below.

Can somebody explain how a 30% tilted angle with less overlap is yelding better GSD than with 90% overlap? Does this have a scientific explanation or is it a software misscalculation?

Thank you very much and have a good week.

[Alexey Pasumansky]

Hello ArnauCM,

How was overlap and GSD calculated in these experiments? Were they taken from Metashape report or orthomosaic meta data, or whether any custom calculation has been applied?

[ArnauCM]

Hello Alexey,

Thanks for your reply. He used Pix4D. My guess is that there is something wrong with how this program calculates the GSD.

However, I wanted to ask if any photogrammetry expert could see a logical, scientific explanation before assuming it is Pix4D's fault.

I might ask him to send me the data and process it with Agisoft to see its GSD values.

Until then, any reasonable explanation?

[Alexey Pasumansky]

Hello ArnauCM,

If it is possible to get the information, how the GSD values from the table has been calculated, it would likely give an answer, why the value is lower for lower overlap. Could be something to do with total number of tie points, tie point projection number or similar.

For example, Ground Resolution value in Metashape report (second page) is calculated in the following way:
at first for each camera GSD is estimated individually as the flying altitude (as average distance from the camera center to the sparse cloud points that have valid 2D projections on the given photos multiplied by the angular resolution of the pixel. Then the value is averaged for all cameras.
Also ground sampling resolution can be calculated as flight altitude in meters divided by the focal length in pixels. The latter can be calculated as focal length in mm divided by the pixel size on the sensor.

So if the datasets are available, projects could be processed in Metashape in order to check the ground resolution values calculated by Metashape for the same tilt angle, but different overlap.