A geometric reasoning approach to hierarchical representation for B-rep model retrieval

•Hierarchy of B-rep model is generated by geometric reasoning automatically.•This shape descriptor eases global and partial retrieval at level of detail.•Interactive prototype system gives why and how similar features are matched.

3D solid model similarity is dependency of many intelligent design applications, such as design reuse, part management, case-based reasoning, and cost estimation. Matching and comparing its intrinsic boundary representation (B-rep) is a critical issue to retrieval. In this paper, we proposed a geometric reasoning approach to generate hierarchy for B-rep model retrieval. We extracted the winged-edge data structure to support series algorithms for underlying geometric reasoning, which is mainly composed of 3 steps to build hierarchy: partitioning, assembling and simplifying. This hierarchical representation is featured with level of detail (LOD)(LOD) retaining geometric and topological information which is proved to be efficient in both global and partial retrieval. Our approach is based on the standard for the exchange of product information (STEP)(STEP), which is suitable for data exchange between heterogeneous CAD systems. The result of case studies from prototype implementation demonstrates its effectiveness and efficiency.