Accurate integration of multi-view range images using k-means clustering
Pattern Recognition
Robust principal axes determination for point-based shapes using least median of squares
Computer-Aided Design
Incremental unsupervised three-dimensional vehicle model learning from video
IEEE Transactions on Intelligent Transportation Systems
| Hi-index | 0.01 |
An automatic 3D model reconstruction technique is presentedto acquire complete 3D models of real objects. Thetechnique is based on novel approaches to pose estimationand integration. Two different poses of an object are usedbecause a single pose often hides some surfaces from arange sensor. The presence of hidden surfaces makes the 3Dmodel reconstructed from any single pose a partial model.Two such partial 3D models are reconstructed for two differentposes of the object using a multi-view 3D modelingtechnique. The two partial 3D models are then registered.Coarse registration is facilitated by a novel pose estimationtechnique between two models. The pose is estimated bymatching a stable tangent plane (STP) of each pose modelwith the base tangent plane (BTP) which is invariant for avision system. The partial models are then integrated to acomplete 3D model based on voxel classification definedinmulti-view integration. Texture mapping is done to obtain aphoto-realistic reconstruction of the object.