Registration of scan data with UAV

[Ingsayyad]

We have scanned as well as imaged an object. I want to register scan and image data together.
At the moment we are using just natural features as tie points between two data sets.
But for higher accuracy, I'm planning to use targets. What sort of targets should I take. Also
shape, size etc.
Flying height: 20-30 m.
Any suggestion??

[stihl]

I'd always go for rectangular or circular shaped targets. Black/white checkered pattern.

For your flying height anything between 10 and 30cm diameter is good enough.

[nadar]

if you print a siemens target, you will also be able to acurately measure the effective resolution of your images.
Add a few colour reference strips on the target, and you will have a simple solution for color calibration.

[Ingsayyad]

@ nadar
Thanx for ur valuable input. But my problem is that I have to locates these targets in my scans  also. 
@ stihl
Can I automatically find these targets ( subpixel) in Agisoft. I need such targets which can be accurately located in scans  as well as in images. Rectangular targets (Black/white checkered pattern) are good for scans but I'm not sure for automatical detection in images.

[nadar]

I never used this feature, but I think you can print special coded targets allowing for automatic recognition.
Have a look to tools/markers/print markers

[stihl]

That's the problem with using un-georeferenced aerial photo's...

It's a hell locating all the GCP's in Photoscan manually. Hence you need the image positions.

[saskiag]

Hi,

I just did some test with different targets (squares, simmens charts, circles etc.). I found out that the best option is to have a circle with a black point in the middle. It is similar of the kind of markers you can print out from Photoscan (in Tool menu). I printed them on a A0 paper sheet and I have a center that has around 15 cm diameter.
We flew at around 115m above ground to have a 4cm pixel resolution with our camera (12MP), and I could find the markers' center with good accuracy in PhotoScan.