Changes to make this more useful as an external library

[slensgra]

I'd like to get the ROS node I have running put into a more finished state by separating out the STag code and the ROS wrapper. To get that node running I had to make a couple of changes to this code:

• There was a memory leak in the EDInterface, so I added a destructor
• The code only compiles on a windows computer unless you remove the windows specific stuff (thankfully not necessary to detect markers!), so I just deleted it (Timer.cpp and the function that finds files in a directory)
• No CMakeLists.txt was provided, so I made one to compile Stag.cpp into a library
• I patched the marker being decoded multiple times problem
• The Stag class has a vector of Marker objects which would increase in size indefinitely over all frames, so I made it clear the markers
• Made detectMarkers return the number of markers detected
• Made the markers vector on the Stag class public to make the markers usable by other code
• Changed the define of NULL to STAG_NULL to avoid a compiler warning
• Put everything into the stag namespace to avoid name collisions in client code
• Added explicit termination criteria into the PoseRefiner.cpp code for the pose optimization

Are you interested in having these changes contributed back into your code? If so, I can do the separation work and improve the style of what I have then make a PR into this code. I could also add an alternate implementation for the windows specific code which would work on linux.

Here is a diff for your reference (messy but can be cleaned if you want the changes).
master...dartmouthrobotics:master

[Norfo]

Hello, @slensgra! Can you explain where you added destructor, please. I'm also faced with this problem.

[bbenligiray]

Hey @slensgra , I'd be happy to merge your PR!

[slensgra]

Great, let me get one ready.

@Norfo See destructor in EDInterface at line 27 of EDInterface.cpp in the diff above

[Norfo]

@slensgra, thanks! By the way, @bbenligiray, do you have any plans to rewrite ED classes? As far as I understand, they are quite old. I faced with a problem that they cannot be used together with OpenCV4. In addition, there is a possibility to rewrite them with cuda support.

[bbenligiray]

The ED code was my coauthors' work and I barely wrapped it for this project. Cihan Topal had a student refactor it with C++ STL (https://github.com/CihanTopal/ED_Lib). As far as I know it's slightly slower, but I would prefer to use that for production. I think there is some ongoing work to integrate this version to OpenCV by third parties.
opencv/opencv_contrib#2313

Especially the edge segment detection part of the algorithm is considerably sequential, so it requires some modifications for a CUDA implementation. There is a proof of concept below from 2012, but I didn't see the code being released.
https://ieeexplore.ieee.org/document/6152450