Title :
A novel image-based method for conveyor belt rip detection
Author_Institution :
Sch. of Autom., Univ. of Electron. Sci. & Technol. of China, Chengdu, China
Abstract :
Conveyor belts of all types are prone to rips from impact by sharp objects in the flow path, which can lead to expensive belt failure. In this paper, we present a novel method for detecting conveyor belt rips. In contrast to previous work on conveyor belt rip detection, the proposed method has a built-in real/false rip classification mechanism, which we believe will serve as a quite useful reference to designers of vision-based conveyor belt monitoring systems who wish to reduce false rip detection rates.
Keywords :
belts; condition monitoring; conveyors; fault diagnosis; image classification; mechanical engineering computing; conveyor belt rip detection; failure; image-based method; rip classification mechanism; vision-based conveyor belt monitoring systems; Belts; Educational institutions; Eigenvalues and eigenfunctions; Feature extraction; Image edge detection; Support vector machine classification; Training; conveyor belt rip detection; digital image processing; pattern recognition;
Conference_Titel :
Signal Processing, Communication and Computing (ICSPCC), 2013 IEEE International Conference on
Conference_Location :
KunMing
DOI :
10.1109/ICSPCC.2013.6663878
