Accuracy of measurements

[Hughsnews]

Hi All,
I am an archaeologist and I am testing photoscan for the creation of ground plans. I just want to make sure I understand one aspect of it.

One test I did was to place a L shaped piece of metal with 3 known distances on it(quite small, like 0.35m and 0.55m). I use the right angle of the markers on the L in order to produce an orthophoto of the area using them as the horizontal and vertical axis, in order to get a perfectly flat birds eye view.

Anyways, I input the distances into photoscan, create scale bars and get a minimal error of a few mm.

Now to test the accuracy of just using these two points on the L piece, I placed a few control groups of marks anywhere from 2-5m away. These were either 1 or 2m in length.

What I want to know is- what is the error produced when just using the L shaped piece with 3 measurements. Obviously as you add more scales, it gets more accurate.

I think if I create markers for my control group points, then create scale bars and enter the distances (1 or 2m), it will provide me with the error. I just want to check that I do not UPDATE this, as if I did it would modify the error with all the scale bars correct?

So just to get the error using the 3 points on the L, enter distances and hit update. Then to get errors of my controls enter distances but do not hit update. Correct?

Thanks in advance for what is clearly an easy question- I just want to make sure.

[gatsri]

Hi

may be it is easier, when you produce an orthophoto, using these L-bar-controlpoints. then check the distances of your checkpoints in an external GIS.
like this, you are sure that you do not influate the model by measuring the checkpoints in your model.

if you have any results, please poste them...