Evaluation of Surface Quality in the Prefabrication of Concrete Panels using Computer Vision

Abstract. The construction industry increasingly uses prefabrication for the significant advantages in cost-effectiveness, production speed, and environmental sustainability. Achieving high-quality building outcomes demands rigorous quality control of the individual building components. Concrete panel manufacturing presents unique challenges, particularly in surface quality assessment, as the material’s heterogeneity complicates standard evaluation metrics from industries like automotive manufacturing, where computer vision methods are well established. This research addresses these challenges with a robust computer vision-based methodology for surface quality assessment for prefabricated concrete panels, focusing on the detection of visible color differences using the ΔE metric. By recognizing the complex material properties and design requirements — including aesthetic aspects like color and texture — the proposed approach establishes more precise and adaptable evaluation techniques to enhance the overall quality and reliability of prefabricated concrete construction.