When it comes to AI, we often hear exaggerated debates about its capabilities, but we shouldn't be worried about these novel predictions overshadowing the truth that AI has emerged and developed over many decades, with many useful advancements predating ChatGPT and its counterparts. There's no better truth than when we consider some practical ways it's being used in production, particularly in quality control (QC).
Machine vision refers to the technology that enables computers to 'see'. It involves the use of cameras, sensors, and specialized algorithms to capture, process, and analyze image data. Although computer vision has traditionally been seen as unrelated to AI, these two technologies are becoming increasingly intertwined as developers turn to neural networks to enhance computer vision algorithms and improve their accuracy.
For example: Inspecting weld quality using machine vision.
AI can assist by enhancing machine vision. And analyzing the root causes stems from the use of artificial intelligence.
Traditional methods include Pareto analysis, fishbone diagrams, along with many other techniques. While these remain essential tools for quality experts, they also demand significant knowledge and expertise to be most effectively utilized. This poses a major issue regarding personnel, concerns about workforce turnover. The proliferation of data then makes deploying artificial intelligence as a natural solution.
Take all your product data and manufacturing processes, then feed them into a machine learning model. Through training, the final model will recognize the correlations between product defects and the causes of those defects. The process is performed continuously, with data accumulating over time, the accumulation of the AI model as an asset of the business.