Reverse Engineering refers to the construction of solid models from existing parts. The process is initiated by taking dense surface measurements using an optical range sensor. These are then approximated to form a surface mesh representation. Polygonal surfaces meshes form the input for solid model construction. However, the C0 continuous meshes are devoid of high level form feature information which conveys design intent or engineering significance. Currently, only partial solutions to detect such features are known in literature. In this research we present algorithms to segment polygonal surface meshes into functional surface features. The first part of the paper deals with simple shape primitives based on 1st and 2nd order surfaces and combinations thereof. In the second part of the paper we propose methods to detect instances of parametric freeform surfaces. Detection is enabled by clustering and pattern recognition techniques. Algorithms and examples are presented. Future research directions are suggested.

This content is only available via PDF.
You do not currently have access to this content.