Computer Vision-based Woven Fabric Defect Detection System

Project ID: WUB/2022/P1/019

Project Duration: 2022 - 2023

Project Leader: Abu Salman Shaikat (Department of Mechatronics Engineering)

Project Members: Rumana Tasnim

Fabric defects are one of the main causes of production losses in the textile industry. This work aims to the development of computer vision-based fabric defect detection systems for textile industries. Different machines are responsible to detect different types of faults in fabric such as blotch or oil spot, broken end, broken pick, crack mark etc. This work introduces an intelligent fabric defect detection system, where only one machine can detect the entire fault of fabrics This system is entirely controlled through the use of image processing techniques. Image Processing helps detect glass defects by using the Haar cascade method. This process is held automatically by a camera, and the upper part of the fabric is checked by this camera. The conveyor belt is used to move the product from one end to another. After detecting the defect, the image will be captured and stored in the system. Moreover, the system will show the type of defect that occurs in the fabric and the conveyor belt stops automatically and the buzzer sounds the alarm. In addition, this data will be recorded in an Excel file for future purposes. Finally, Fabric defect data will be collected to prove the effectiveness of the proposed system. This system will support as an extremely effective fabric defect application.