چكيده به لاتين
One the main operations in the reverse engineering of mechanical parts such as turbine blade is surface reconstruction using points cloud which are concluded from all the data capturing methods. On this base using points cloud for creating 2D profile of sections in order to remake special surface features like extrude surfaces and sweep surfaces from the view of final surface geometry consideration and also stability of results from administrative process, has much importance. Moreover, presentation of an automatic framework which is able to beside of accelerate reaching process to CAD model, prepare the groundwork for its design requirement, will lead to decrease leading time of this product. Therefore, in this research, it has tried to discuss that the most efficient creation action of sectional 2D profile in a feature-based framework using constraint fitting, moreover to provide a situation that by using an automatic geometric feature recognition attain to the nearest CAD model in comparison to design intents. This framework divide to four parts as points cloud preprocessing, segmentation, classification and final fitting. In order to perform the said parts have been using from concepts like ML clustering, CPD, RANSAC, binary tree with inorder traversal, automatic constraint setting and constrained/unconstrained optimization.
