Surface Completion for Scanned Heritage Sites
Proposer: James Gain
Brief Description: It is common practice in Geomatics to digitally capture the shape of buildings for analysis purposes, using laser range scanning. This produces a cloud of points that represent the position and colouration of samples on the building surface. Unfortunately, due to constraints on the position of the laser scanner and the convolutions of the building these point clouds often have gaps. An automated approach to filling these holes is to transplant complete surfaces from elsewhere in the model, using the edges around the hole as context.
This project will involve implementing a context-based surface completion algorithm  and integrating it into an existing OpenSource point-based modelling system . Development will be done in collaboration with the Geomatics Department at UCT, which will provide real-world data from heritage sites in Africa, such as the stone churches in Lalibela, Ethiopia.
 Sharf, A.,, Alexa, M. and Cohen-Or, D.  “Context-based Surface Completion”, In Proceedings of SIGGRAPH ‘04, pp. 878-887.
 XGRT graphics toolkit: http://www.mpi-inf.mpg.de/~mwand/XGRT/index.html
Computer Science Content: Computer Graphics (point-based modelling, hierarchical data structures)
Specific Learning Outcomes: Effective project integration, Advanced computer graphics algorithms
Skills Required by Team as a Whole:
Theory: moderate [understanding advanced computer graphics algorithms]
Implementation: moderate to hard [integration with an existing OpenSource system and implementation of spatial partitioning hierarchies and cell-based matching]
Hardware: dual-core PC with reasonably up-to-date GPU
Software: C++ IDE and OpenGL, OpenSource XGRT system
Data: laser scanned heritage data-sets provided by the Geomatic department
Supervision: James Gain and Heinz Rüther (Geomatics), weekly meetings with supervisors, occassional attendance and presentation at research group meetings may be required.
Number of Students: 2 or 3