Skip to main content
Disciplines
Automotive Business IT + Informatics Construction + Real Estate Electrical Engineering + Electronics Energy + Sustainability Insurance + Risk Finance + Banking Management + Leadership Marketing + Sales Mechanical Engineering + Materials
Events
DE
EN
Books
Journals
Topic Page
Marketing
Start single access now
Access for companies
EXTENDED SEARCH
Log in
Top

2020 | Book

Introduction Read chapter Read first chapter

Machine Learning Methods for Reverse Engineering of Defective Structured Surfaces

Author: Pascal Laube

Publisher: Springer Fachmedien Wiesbaden

Book Series : Schriftenreihe der Institute für Systemdynamik (ISD) und optische Systeme (IOS)

Part of: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Login to get access
insite
SEARCH

About this book

Pascal Laube presents machine learning approaches for three key problems of reverse engineering of defective structured surfaces: parametrization of curves and surfaces, geometric primitive classification and inpainting of high-resolution textures. The proposed methods aim to improve the reconstruction quality while further automating the process. The contributions demonstrate that machine learning can be a viable part of the CAD reverse engineering pipeline.

Table of Contents

Frontmatter
Chapter 1. Introduction
Abstract
Modern functional plastic surfaces are usually produced using injection moulds. In plastic injection moulding, the respective material is plasticized and injected under pressure into an injection mould made of hardened steel. After hardening to a solid state, the moulded part can be removed from the injection mould.
Pascal Laube
Chapter 2. Fundamentals
Abstract
In this thesis we will consider two main areas of application for machine learning namely classification and regression. In the so-called training phase the machine learning algorithms parameters need to be adapted with respect to the problem. Let us define a set of n training samples xi with class labels yi together called the training set.
Pascal Laube
Chapter 3. Parametrization in Curve and Surface Approximation
Abstract
For the approximation of B-spline or T-spline surfaces to point clouds the computation of a parametrization is required. Since the construction of both surface types is based upon the concept of B-spline curves, we begin by developing parametrization methods for curves. Parametrization for B-Spline curve approximation includes the computation of a point parametrization t as well as the knot vector u for a given set of ordered points.
Pascal Laube
Chapter 4. Classification of Geometric Primitives in Point Clouds
Abstract
As introduced in Section 1.1 classification of geometric primitives in point clouds is an integral part of the reverse engineering pipeline. It is also a vital part in the particular case of separating surface structure from base geometry. Compared to object recognition or object retrieval tasks the classification of geometric primitives is somewhat harder.
Pascal Laube
Chapter 5. CNN Texture Synthesis for High-Resolution Image Inpainting
Abstract
In most modern manufacturing processes surface structure is stored separate from the basic geometry of the surface. A suitable counterpart from computer graphics are displacement maps which store a displacement on actual surface points to enrich models with detail. While displace-ment maps can have arbitrary depth resolution they are often represented as 8-bit grayscale images.
Pascal Laube
Chapter 6. Concluding Remarks & Outlook
Abstract
In this thesis, I presented novel machine learning approaches for key problems of the reverse engineering process of structured surfaces. The developed methods aim to further automate the steps of reverse engineering further. A particular emphasis is laid on the reconstruction of the surface’s base geometry as well as surface structure which is mandatory for many functional surfaces.
Pascal Laube
Backmatter
Metadata
Title
Machine Learning Methods for Reverse Engineering of Defective Structured Surfaces
Author
Pascal Laube
Copyright Year
2020
Electronic ISBN
978-3-658-29017-7
Print ISBN
978-3-658-29016-0
DOI
https://doi.org/10.1007/978-3-658-29017-7

Premium Partner

    Image Credits