User-Driven Computer-Assisted Reverse Engineering of Editable CAD Assembly Models

Abstract

This paper introduces a novel reverse engineering technique for the reconstruction of editable CAD models of mechanical parts' assemblies. The input is a point cloud of a mechanical parts' assembly that has been acquired as a whole, i.e. without disassembling it prior to its digitization. The proposed framework allows for the reconstruction of the parametric CAD assembly model through a multi-step reconstruction and fitting approach. It is modular and it supports various exploitation scenarios depending on the available data and starting point. It also handles incomplete datasets. The reconstruction process starts from roughly sketched and parameterized geometries (i.e 2D sketches, 3D parts or assemblies) that are then used as input of a simulated annealing-based fitting algorithm, which minimizes the deviation between the point cloud and the reconstructed geometries. The coherence of the CAD models is maintained by a CAD modeler that performs the updates and satisfies the geometric constraints as the fitting process goes on. The optimization process leverages a two-level filtering technique able to capture and manage the boundaries of the geometries inside the overall point cloud in order to allow for local fitting and interfaces detection. It is a user-driven approach where the user decides what are the most suitable steps and sequence to operate. It has been tested and validated on both real scanned point clouds and as-scanned virtually generated point clouds incorporating several artifacts that would appear with real acquisition devices.

Case-based tuning of a metaheuristic algorithm exploiting sensitivity analysis and design of experiments for reverse engineering applications

“…It also handles the tasks of performing the sensitivity analysis of parameters and their grouping using K-mean clustering technique. For the ease of users, a prototype software created in VB script has been integrated as a plugin in SolidWorks® 2017 Education Edition [5] that allows efficient implementation of the proposed fitting technique.…”

“…Distance computation is performed using CloudCompare called in batch mode to compute the nearest triangle distance against each point in the point cloud. Alternatively, point-to-point distance can also be used as an energy function if the tessellated mesh is sampled with points [5]. As detailed in [4], this process is further improved while allowing points of the PC to be filtered step after step, so as to allow local fitting of a part in the point cloud of a digitized mechanical assembly.…”

“…There are few methods to reconstruct editable CAD geometries. Among them, metaheuristic algorithms like Simulated Annealing (SA), Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) have been cleverly used to optimize the parameters of the CAD geometries to fully fit in the point cloud obtained from data acquisition devices [4][5][6][7]. These metaheuristic algorithms try to find global optimal solutions by both diversifying and intensifying the search in a large solution space.…”

Explicit Annotated 3d-CNN Deep Learning of Geometric Primitives Instances

“…The improved recognition rates of artificial neural networks in computer vision tasks satisfy these needs with convolutional architectures in a deep learning (DL) manner (Bici et al, 2020). An enhanced computer-based geometric recognition and detection in RE with additional information for subsequent algorithmic processes would promote a more coherent assessment of the implicit parametrization by the technical components RE expert (Shah et al, 2022). Although many geometric detection approaches apply DL to the process step of surface reconstruction, other methods that allow a deeper geometric model understanding in the sense of inspection techniques for surface defects are yet to be known (Geng et al, 2022).…”