Is there any progress about an SDK allowing to retrieve some information from the device in real time?
I have objects with special markers. I can detect the markers in real time. The markers are some kind of millimeter grids.
I want to scan these objects. I need to have an 3d object with a texture containing these markers. If I am using the final texture made by the scanner, the quality is not enough accurate and I have several pieces which break my algorithm.
If I can access to the current frame video, I will not be dependent on the final quality of the texture or to the unfolding algorithm.
I would need to access additional information:
- The current image frame to detect some data with my own algorithm in real time.
I have an algorithm that detect my special markers on my objects. I can find the matches between frames.
With this algorithm, I am able to compute the rotation, translation between two frames that can be used to feed the registration algorithm.
It also allows me to remesh the model
- The current point cloud generated by the device for each frame
- For each registered point cloud, the transformation matrix associated to it
- The final 3D mesh