SMYG – Smart image analysis for surface defects on cast components

Swedish aluminum, iron, and steel foundries want to improve their ability to detect and assess defects and deviations on the surface of cast components, to ensure consistent and high quality of delivered products. The working environment and the content of the quality control work need improvement, as it currently involves physically demanding manual handling with heavy lifting, and significant responsibility under often high time pressure. The foundries also want to reduce scrap and remelting of materials and components, thereby becoming more resource-efficient in terms of material use and energy consumption. The idea of the project is therefore to streamline the ongoing quality control of castings using computer vision systems and physics-based machine learning.

The goal is to pave the way for a digital computer vision tool for the next generation of quality control of surface defects on cast components. Today, the use of computer vision in  foundries faces two major challenges: the lack of relevant training data and the demanding industrial environment in a component foundry. This project will develop hybrid methods with physics-based machine learning to produce relevant synthetic training data for the vision systems and establish guidelines for the design and requirements of a future digital tool. In the long run, this will enable new and combined data flows between different systems within foundry production for new and expanded applications.