Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.341288
Title: Enhancements to reverse engineering : surface modelling and segmentation of CMM data
Author: Bardell, Rayman A.
ISNI:       0000 0001 3444 6527
Awarding Body: Nottingham Trent University
Current Institution: Nottingham Trent University
Date of Award: 2000
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Reverse Engineering (RE) is a relatively new field, which aims to reproduce a physical prototype accurately. This involves the main stages of data collection, modelling and machining. One approach to data collection utilises the Co-ordinate Measuring Machine (CMM), using a touch-trigger probe. This contact measurement method is employed in this work, resulting in three-dimensional (3D) surface data of a physical part. Surface modelling follows data collection as the next stage of RE, where a best-fit model is generated, interpolating these data points. An automatic surface generation methodology is developed in this work, utilising the ACTS® Computer Aided Geometric Design (CAGD) development tool. This results in C* continuous Gregory/Charrot patches, which are merged to form a global B-spline surface model. Classic free-form modelling techniques have associated limitations, affecting accuracy. This work aims to reduce these errors, improving the accuracy of the RE process. The initial study has identified two types of global surface in physical prototypes, known as free-form continuous surfaces, and composite discontinuous surfaces. These require different modelling strategies to minimise modelling errors. There has been limited research in the area of surface model analysis, applied to automatically generated surfaces. This present work focuses on a novel method for analysing the accuracy of a generated surface model, highlighting regions of the surface that deviate from the CMM data. With free-fonn objects, deviations can be used to define the smoothness of the CMM data. With composite objects, these deviations occur at areas of extreme curvature change. The development of a data-set partitioning tool investigates a unique method of predicting surface model accuracy, where a model is generated from a fraction of the available data-set. The accurate modelling of composite surfaces derived from CMM data has received little attention to date. In this work, surface models of this type are interrogated, using a novel curvature-based surface decomposition method, allowing individual smooth local sub-surfaces to be modelled. This utilises a seed region growing methodology, which clusters points of the same surface type as a sub-surface, allowing assessment of the sub-surface in terms of cosmetic quality. This provides a method of determining and segmenting specific surface types from the global surface, reducing inaccuracies caused by the modelling of adjacent surfaces of differing types. The main areas of novel research presented in this thesis lie in the areas of accuracy analysis, introducing deviation analysis, and surface type recognition, developing a novel seed region growing methodology, applied to CMM data.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.341288  DOI: Not available
Keywords: Co-ordinate measuring machine; Computer aided
Share: