Projector-camera system is used for implementing spatial augmented reality which project virtual contents onto real object and 3D scanning. In this system, a gray code structured light sequence is used for encoding projector coordinate because a proje...
Projector-camera system is used for implementing spatial augmented reality which project virtual contents onto real object and 3D scanning. In this system, a gray code structured light sequence is used for encoding projector coordinate because a projector doesn’t take a picture like camera. And then, we set projectorcamera correspondence after decoding encoded projector coordinates for using captured images. However, we found incorrect set of projector-camera correspondence caused by effect of ambient light and strong interreflection from other surface patches. It leads to a result of incorrect 3D scanning. In this paper, we propose a method to improve projector-camera correspondence by using sub-pixel classification. Our method will improve projector-camera correspondence after classifying correct correspondence and incorrect correspondence automatically. The proposed method has advantages robust to varying threshold values in contrast to previous methods. And, it can be used to a variety of applications.