Abstract The detection of defects in laminate flooring using machine vision based on neural networks an‎d artificial intelligence has been conducted in this thesis. Laminate flooring is a type of wood flooring consisting of four layers: wear, decor, core, an‎d moisture-resistant overlay. Finding defects in laminate flooring can be very difficult an‎d time-consuming due to the large surface area of the flooring, small dimensions of defects, an‎d intricate background patterns. The most important challenge in finding defects in laminate flooring is detecting small defects in a timely manner, as failure to do so can lead to accumulation of defective products in the factory. Currently, the inspection of these defects in factories is done manually, which requires a significant amount of human resources due to the high production speed an‎d small size of defects. The aim of this research is to develop a quality control system with innovation in processing classification technique using ceramic coating to identify defects in minimal time, determine their location an‎d quantity in a very short time, an‎d transfer this information to the employer. The stages of this research are as follows: firstly, based on previous models an‎d experiences, a new an‎d optimized model that can identify all desired defects with minimal equipment has been designed in Inventor software. Then, using AI techniques, all of these defects were trained to be recognized by the system immediately after production on a sample of 340 laminate flooring. Based on the results obtained in this thesis, it can be concluded that this method is a complete an‎d practical approach for a real production line, with an error rate close to zero an‎d the capability of updating, optimizing, an‎d customizing based on production characteristics.

حامد رضوي بني , عليرضا مومني

بهمن اسدي , محمداسماعيل نظري