Description
Today a lot of buildings, cultural heritage sites, caves, ground strips, to name but a few, are recorded by distance measurement devices like laser scanners. The result of such recordings are point clouds consisting of millions or even billions of points. Converting such points clouds to triangle meshes is a challenging task, as the point clouds are often not complete, which is difficult for meshing algorithms to handle.
Task
Implementation of a complete processing pipeline from point clouds to triangle meshes, from the original laser scan data to the complete mesh of a scanned object. This includes converting the point data to a triangle mesh and then rendering the polygonal mesh. The point data is provided divided up into several smaller files, which have to be processed individually and then merged into a single large mesh eventually. Rendering the resulting large mesh is not the main topic. A renderer for visualizing the mesh should be implemented with the help of
OpenSceneGraph.
Requirements
The students should be interested in the following topics: point cloud to mesh conversion, large data sets, mesh processing. Good C++ knowledge is mandatory.
Environment
The environment will be a rendering framework that is developed at the institute, named Scanopy. It is targeted for processing and rendering large point clouds and shall be expanded for processing and rendering meshes as well. The programming language is C++, the UI is based on the Qt framework. Development will be with Visual Studio 2010, 32-bit or 64-bit.
