I'm struggling to understand how to solve the following problem:
Given
- Input image that has some homes in it
- Camera calibration: 3X4 projection matrix
- 3D scan in the following format: [X coordinate, Y coordinate, Z coordinate, R value, G value, B value, X normal, Y normal, Z normal]
- Marker image I want to put on particular home (for example like a 'welcome' sign or for sale)
- 4 coordinates in 3d where to put that image
I need to generate a new image w/ only the home I put my marker image on, and I can't just use hand-coded values from the 4 coordinates in 3d I'm given, I need to calculate everything during runtime. (Camera calibration and 3d scan in the same world coordinates)
I'm not sure how to even approach this, I think that it should be something similar on how markers are done in augmented reality (that's just my guess).
Any help would be much appreciated.
Input Image for example:
Output Image:
