georeferencing accuracy

[André Jalobeanu]

Hi,

I’m having some accuracy questions about the anchor and target based georeferencing when a projection is used (eg. UTM)
Has anyone tested the accuracy of this linear model and the practical consequences?

 
In theory, for each emitted pulse there is a train of returned waveforms that are close to each other 

(50-100 m usual, maybe a few 100 m for tall buildings, Eiffel tower etc.) so it should not a big problem to use a linear approximation.
But in practice there are air waves, birds artifacts etc. so the actual range of returned waves might be very long in some cases. 
Wouldn’t that compromise the accuracy of georeferencing?

 

I ran a few simple experiments, and a 3000 m range at 30 deg from vertical in UTM already shows significant errors 

(anchor point is laser origin range=0, target point range=3000, max horizontal error range=1500)

A simple solution would be to duplicate the pulse record and separate the wave blocks if they are too far apart, 

but that's not compatible with the format specs... 

Any thoughts? 

André