Novel textile imaging artificial Neural network techniques for Plain woven fabric defect Identification

Abstract

The objective of this research is to automate Fabric Inspection newlineprocess by a computerized model based on Image Processing using Artificial newlineNeural Network techniques Obviously fabric inspection plays a crucial role newlinein textile industry as it prevents the risk of delivering inferior quality products newlineto its customers Till date conventional fabric defect identification is carried newlineout in offline and defects are identified manually with many drawbacks such newlineas less accuracy, time consuming and expensive Usually the fabric after newlineproduction is doffed in large rolls from weaving machines and dispatched to newlinethe fabric inspection department newlineA skilled human inspector places the rolls on the inspection newlinemachine with sufficient lighting to identify defects In this kind of process the newlinetextile industry faces a lot of challenges in delivering high quality fabric newlinebecause of its slow processing speed lack of skilled inspectors high newlineproduction cost and failure to produce 100 defect free fabric Therefore, newlineautomated fabric defect identification is need of the hour in textile industry newlineSince an automated fabric defect identification system is composed of a newlinecomputer system with defect identification software it does not suffer from newlinethe drawbacks of conventional fabric inspection system newline newline