Where mesh.obj obtained using “Scanner - Structure Sensor Sample”. Given set of points (u,v) on the image find corresponding mesh vertices (X,Y,X).
For that I made some changes in the “Scanner - Structure Sensor Sample” code and add a button that saves rgb-image and cameraViewpoint matrix each time I’m pressing it.
To visualize it I used “c++” and “cv::viz::Viz3d” class.
So for example I took two images and there cameraViewpoint matricies during scanning of the object. Visualized it and draw a line from origin of each camera to some point on the mesh.
I would expect that the line will pass through same point on the image on two images and they will correspond to 3d mesh point.