Hi everyone,
I'm Vincenzo, I'm trying to automate the process of face scanning through the use of 14 Raspberry Pi cameras. After several attempts I was confronted with the following problems:
Problem 1.
I made a base project of face scanning who got good results (Chunk 0) and by this I start to analize the following three cases:
Case 1) If I save the camera's caibration certificate of the Chunk 0 and then create a new Chunk (Chunk 1) adding a new set of photo and importing the calibration certificate of Chunk 0, I get poor results (see Image 1)
Case 2) If I export the camera's calibration certificate of the Chunk 0 from Tools -> Export, and then I create again a new Chunk (Chunk 2) with the same certificate and a new path of photo, i get poor results again (see Image 2)
Case 3) If I copy the Chunk 0 and then I change the path of photo from the multiframe layout at the end i get the best results of the three cases (see Image 3)
Someone can explain me why all this happens and how it's work?
Problem 2.
The second problem is streactly connected with the case 3 above. When I change the photo path from the multiframe layout of the copied Chunk why the Preview didn't update? There is same way to resolve this problem, maybe with some xml script?
Problem 3.
Since I want to automate the process, there is a way to insert in a copied chunk the marker points of an old one with the Batch Process?
All your comments are very helpful! Thank you!
Topic: Face scanning problems (Read 3920 times)
