IMProVe 2011 - Proceedings
IMProVe 2011 - Proceedings
IMProVe 2011 - Proceedings
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Image Processing and Analysis<br />
carry out manually the edge dyeing procedure represents a bottleneck for the entire<br />
leather finishing process. As a consequence the development of an automated dyeing<br />
procedure is advisable. The main purpose of the present work is to describe a Machine<br />
Vision based system for automatically dyeing leather patch edges. Such a system<br />
comprises two main parts: a Machine Vision (MV) system devoted to leathers edge<br />
detection and a CNC painting pantograph (2 axis Cartesian robot) whose painting tool is<br />
moved according to the output of the MV apparatus. The developed system, tested on a<br />
prototype dyeing machine, proved to be effective in delivering high quality edge finishing.<br />
Keywords: Leather dyeing, Machine Vision, Process automation.<br />
Corresponding Author: Lapo Governi<br />
Tel.: +39-554796396<br />
Fax.: +39 -554796700<br />
e-mail: lapo.governi@unifi.it<br />
Address: via di Santa Marta, 3, 50139 Firenze.<br />
Automatic image-based car profile character line recognition<br />
M. Fiorentino (a), A. E. Uva (a), G. Monno (a)<br />
(a) Dipartimento di Ingegneria Meccanica e Gestionale, DiMEG, Politecnico di Bari<br />
Abstract:<br />
Purpose:<br />
This paper presents an automatic image based tool for character lines recognition of<br />
automotive profiles.<br />
Method:<br />
A novel Matlab program scans automotive pictures on the web recognizing the wheels<br />
and the main feature lines.<br />
Result:<br />
The system can scan a large dataset of existing or concept car models using image search<br />
engines.<br />
Discussion & Conclusion:<br />
Taking advantage of the formal topology of automotive profiles the wheels are used as<br />
main reference for character lines extraction. The algorithm detects wheelbase, roof,<br />
waist and accent line with a good level of confidence. The collected data can be used for<br />
statistical studies and to define the character of a car design.<br />
Keywords: Aesthetic design, character lines recognition, image processing, computer<br />
aided styling.<br />
Corresponding Author: Michele Fiorentino<br />
Tel.: 00390805962800<br />
Fax.: 00390805962777<br />
June 15 th – 17 th , <strong>2011</strong>, Venice, Italy<br />
67<br />
<strong>IMProVe</strong> <strong>2011</strong> - <strong>Proceedings</strong>