A basic augmented reality card based application in Python. This is a fork of juangallostra's repo.
A sample output
When I first started learning OpenCV/Computer Vision and was looking for ideas to implement and/or implement OpenCV, I made the connection to Augmented Reality. And AR is something that has always fascinated me. So, when I realised I was in fact learning the know-how for AR, I just had to test it out.
So, when I first started out, little did I know that I had loads more to learn before I could "jump in" 😅. But I have to admit, this was a good starting point an interesting project to delve into, it taught me about keypoint feature extraction and orb matching.
The next area I have yet to figure out is the 3D model loading; I haven't touched that code yet.... I tried playing around with the existing code to load textures and/or materials, with no success, yet....
Updates
- Code works on Python2 and Python3 now
- updated parameters in keypoint matching for improved accuracy. And a better filter for good matches. The updates and testing can be found in
orbmatch.pyandorbmatch_video.py
Update: This should no longer be an issue after this commit
See this blog entries for an in-depth explanation of the logic behind the code:
- RD Milligan
- 3D Models
- [Filter]