Cheers,
I am developing a photogrammetry application and I will likely be using one of the commercial packages (like Photomodeler or Acute3D). I believe all of these programs come with a capability for automatic target detection.
The user prints out the supplied pdf targets (at the scale he chooses) and places the targets in the scene. The user then captures the necessary pictures.Software processes the pictures into 3D. As part of this process, the software automatically recognizes the targets which improves the 3D quality.
For various reasons, I would like to create my own targets(or Ground Control Points). This is possible as the Photogrammetry software packages I'l be using allow a user to specify the location of user defined targets (also called Ground Control Points) through an API. To do this, for each Ground Control Point and photo containing a Ground Control Point, the user specifies the pixel location of the Ground Control Point.
I should also mention that the photos are captured at close range (within 5 inches or so). I would expect the targets to be located at sub mm accuracy.
Is there a recommended approach to creating and recognizing targets? I was thinking of using circles with rotating crosses in the middle. I definitely do not want to reinvent the wheel. If there are libraries available to do so, that would be great. What about AR markers or QR codes? Do they provide the capability I need?
Any thoughts/ideas are gratefully appreciated.
Graham