Recovery of three-dimensional objects from single two-dimensional sketches using cubic corners

This report describes the work done to recover 3D objects from a single 2D line drawing using theconcept of cubic corner. The cubic corner approach was adopted because it provided faster, less complex and mathematically accurate recovery of simple 3D models, and a platform to apply the logic...

Full description

Saved in:
Bibliographic Details
Main Author: Vishnuprasad Venugopalapanicker
Other Authors: Lee Yong Tsui
Format: Final Year Project
Language:English
Published: 2011
Subjects:
Online Access:http://hdl.handle.net/10356/42838
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:This report describes the work done to recover 3D objects from a single 2D line drawing using theconcept of cubic corner. The cubic corner approach was adopted because it provided faster, less complex and mathematically accurate recovery of simple 3D models, and a platform to apply the logic in the recovery of more complex shapes if a cubic corner was present. The 3D object recovery from a single line drawing began by identifying the 3D faces. Then the cubic corner was identified in the sketch and its depth was assumed zero. Cubic corner theory was applied to this corner, which enabled to find the real value of its three adjacent vertices. The rest of the vertices in the model were recovered using simple geometrical mathematics. The solution was developed using AutoCAD as the drafting tool and VB.NET as the programming platform. The solution initially catered for planar polyhedral objects. It was later expanded to include polyhedral objects with circular curved surfaces that had a planar axis. Experimental results provided shows recovery is possible for any object in the above class, provided they have a cubic corner. Results were accurate when the input line drawings were accurate. The solution was also fast and simple, meeting the design objectives. Handling of inaccurate line drawings and imprecise cubic corners needed improvement. Recommendations for such have been included. Lack of support for automated face recognition and additional geometric types were identified as shortcomings, and presented as future research topics.