itle: Automatic Detection of Defects in Manufacturing using AI techniques

Abstract: As competition and demand in manufacturing increases, faster
production systems are needed. In all industries and manufacturing areas,
quality assurance and defect free products are vital to assure the high
quality of the products and avoid customer disappointment. Traditionally
quality inspection is performed by human inspectors who either inspect
each and every product individually or inspect samples from the whole
batch. These methods are however time consuming, tedious, costly and still
leave room for errors. However as machine vision advances, automated
inspection is taking over. Many automated techniques have been proposed
and implemented on manufacturing lines such as visual inspection using
cameras, x-ray, laser, etc. The area is purely a combination of Image
Processing and Machine Learning as most of the techniques pre-process
images prior to inspection to enhance the defects and make the detection
process easier. Image pre-processing techniques used include histogram
processing, edge detection, noise removal, etc. On the other hand, defects
are then detected using techniques such as Artificial Neural Networks,
Image Subtraction, properties of light, etc.