TEXTILE CLOTH FAULT DETECTION

Abstract

Textile Industrial Automation aims to take a step towards the better industrial state by targeting the industries with little or no know how of automation and guiding them towards the better output of their stock in the world. With the development in the industrial world, various processes have been automated to make the life of the people more comfortable and this automation is used by the industries now a days frequently. Currently in Pakistan, no such automated processes are available in our industries to provide quick solutions to the industrial problems to fix the declining economic growth of industries in Pakistan as we still rely on the manually operated working in our industries. Traditionally quality inspection in industry takes place via workers which is inefficient and time taking. In this project, we want to automate the quality inspection using industrial cameras and image processing algorithms. We used Faster R-CNN in this project for detection of faults in fabric wit VGG16 architecture. Training consists of around 3600 images and 7 faults are trained with around 95 percent accuracy