Machine Vision in Industrial Quality Control

Abstract

With the increasing requirement of improving productivity and precision, machine vision has gained more and more traction and has continually proven to be an effective method of automated visual inspection. From food to medical equipment to automotive parts, machine vision is being adapted more widely in different industries. With the recent advancements in Artificial Intelligence, Machine Vision has reached a new height, becoming more robust and accurate. Nevertheless, some companies may find it hard to employ this method effectively to their quality control, having to integrate a Machine Vision system to an established control cycle.

The objectives of this thesis were to explore the usage of machine vision in manufacturing and quality control by analysing theoretical background on the topic and discuss how to best create a machine vision system for this purpose. To achieve these goals, the theoretical framework will cover every step when building a machine vision system: acquisition, pre-processing and processing, giving a broad view of the different key parts and their implementation. This is followed by a description of LabView, the programming interface of choice, and its vision-specific package. Finally, a proof-of-concept machine vision system was implemented to demonstrate the feasibility of its inclusion into an existing production process.

Based on the studies presented in this paper, it can be concluded that the application of machine vision is a viable and cost-efficient option for the purpose of industrial quality control. The envisioned objectives were met and practical suggestions on the applicability of machine vision in inspection and quality control are provided. The thesis also provides a guideline to improve upon the proof of concept unto more applications