3D reconstruction of objects has been an active area of research for over 30 years now. In this project we explore the possibilities for 3D reconstruction of rigid objects using depth sensors coupled with traditional RGB cameras. More precisely, we use the RGB feed to find correspondences using the ASIFT (Affine scale-invariant feature transform) algorithm and ICP (iterative closest point) on the depth feed to estimate the transformation between the frames. We finalize the reconstruction by using a level-set reconstruction method to smooth the result and fill potential holes in the accumulated depth map.
Link to this comment:
All Comments (0)