Design recovery of internal and external features for mechanical components

Contemporary reverse engineering (RE) tools focus on generation of free-form shapes from point cloud data collected from scanning systems. The final model contains a set of surfaces and curves that have no functional meaning, and noise due to manufacturing variations or wear are contained within the model. A different approach is required in order to create a more suitable model for engineered components because of these issues. To meet these challenges, a systematic approach is adopted in a comprehensive manner to extract the relevant information and transform it into pertinent design knowledge. A modular design recovery framework is presented that captures the component's structure, function and feature information at varying perspectives. To complement the framework, form recovery algorithms have been developed to transform point cloud data into wire frame geometry consisting of standard line and arc elements. Once the points are converted into curve primitives, adjustments are made to capture the design intent using heuristics, and common shapes and two-dimensional (2D) patterns are detected. From this geometry, a surface or solid model can be constructed using established geometry creation tools. Several case studies are presented that illustrate the form recovery algorithms to highlight their merits.