Toronto Metropolitan University
Browse

Feature Recognition in Geometric Reverse Engineering

Download (3.58 MB)
thesis
posted on 2021-05-22, 17:23 authored by Muhammad Arshad
An artificial neural network based feature extraction system for finding three dimensional features from physical objects is presented. As part of a geometric reverse engineering system, the feed-forward neural network allows for the efficient implementation of feature recognition. Reverse engineering of mechanical parts is the process of obtaining a geometric CAD model from the measurements of an existing artefact. Ideally, the reverse engineering system would automatically segment the cloud data into constituent surface patches and produce an accurate solid model. In order to accomplish this intent, a neural network is used to search and find the features in the initial scan data set. In this work, feature extraction for geometric reverse engineering has been accomplished. Work has also been done to extract features from the multiple shapes. The technique developed will reduce the time and effort required to extract features from scanned data of a physical object.

History

Language

English

Degree

  • Master of Engineering

Program

  • Mechanical and Industrial Engineering

Granting Institution

Ryerson University

LAC Thesis Type

  • Thesis

Year

2004

Usage metrics

    Mechanical and Industrial Engineering (Theses)

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC