number of GCP

[Piwo]

Hello everyone,

I'm working in the "Laboratoire Magmas et Volcan" in France. I make a training period for finish a land surveryor engineer school. I have to make DEM from photography take by plane, drone and satellite.

I want to explain in my dissertation how many GCP i need to make the DTM.
For the drone, i get 900 photography but the drone didn't get a GPS and a inertial reference systems (IRS).
For the plane, i get 44 photography and there are a GPS and a IRS onboard.

My question is : how i can estimate the number of GCP (minimum) that i need to make the DTM. Which parametrizes Agisoft needs to align photography to be able to make an assessment unknown / observation.

Sorry for my poor english, and thanks to take the time to think about my question.

[frank.stremke]

hi
the number of gcp depends on many variables (of course it would be to easy otherwise)
things to keep in mind is overall geometry of the object elevation differences and steps in terain then the resolution and acuracy reqirements as well as the technique for getting GCP
given all that i still say take as many as you need and as few as you have to :-)
usualy they say 10 is a good number which i can agree to if the area is not to big but take some check points as well (leave them unchecked in the ground control panel to allow for error checks)
frank

[stihl]

Piwo,

From experience I've gathered that GCP's can easily be 150-200m apart from eachother.
I've analyzed that if you go further than that there's a chance that the model between the GCP's will concave.

[Piwo]

thanks for the precision.

At school we had seen that for a flight without GPS or IRS, it was needed approximately 1 point know in planimetry every 5/6 couple of images ans 1 point know in altimetry on all the couples. But i think that this estimation is appropriate to the software that we used at school.

Now I use Agisoft and I wanted to estimate(esteem), thanks to the equation of colinearity, the number of GCP by making a balance unknown / observation. That is why I would like to know the parameters (camera, GCP, etc.) used by agisoft to create a dense cloud from photography.