Research Group on Visual Computation

Filling Missing Parts of a 3D Mesh by Fusion of Incomplete 3D Data

Icon of project

Description

Vision is a natural sensor for 3D object reconstruction using Structure from Motion (SfM) techniques. The reconstruction of the 3D scene is possible up to an overall scale ambiguity. The input of SfM methods is a set of 2D images covering the whole object surface, which is easily obtained with modern cameras. Range sensors (e.g. Lidar) provide 3D structure directly, but such devices are more expensive and require careful setup in order to obtain a full metric scan of an object. When using 3D Lidar technology, it is possible to obtain a metric 3D model of an object by aligning scans from different viewpoints. However, it is often difficult to achieve a full coverage of the object because of the limitations in viewpoints, yielding a partial metric model.

We propose a novel algorithm which uses two datasets to produce a complete and accurate 3D model by filling in missing parts (holes) of an accurate reference 3D model using a less accurate but more complete moving 3D model. It operates only on a small region around the holes and computes a locally non-linear transformation which moves the less precise surface into the hole such that local structural properties are respected but point-wise noise is removed.

The method is quantitatively evaluated on a real dataset, which confirms its performance both in terms of quality and computational efficiency.

Results

Textured visualization of a planar sample (top) and the relief sample (bottom). From left to right: the original Lidar data; the original patch of the moving model; the Lidar data with processed moving data. Note that the geometry of the patch is kept intact.

Publications to cite:
  1. Laszlo Kormoczi, Zoltan Kato, Filling Missing Parts of a 3D Mesh by Fusion of Incomplete 3D Data, In Proceedings of Advanced Concepts for Intelligent Vision Systems, Springer, Antwerp, Belgium, 2017. [bibtex]

Hichem Abdellali has been awarded the Doctor of Philosophy (PhD.) degree...

2022-04-30


Hichem Abdellali has been awarded the KÉPAF Kuba Attila prize...

2021-06-24