Managing waste effectively is very important in industries to save resources and reduce costs.  Traditionally, the identification of waste has relied on manual inspection, which is time-consuming, costly, and susceptible to human error. This is especially problematic in factories where production lines change often, making manual waste checks even more costly. In response to these issues, this paper introduces a novel approach aimed at automating the process of waste detection. We propose a structured repository of rules designed to categorize industrial activities into eight distinct waste categories. The development and validation of these rules are detailed, and their implementation has been tested in the Fischertechnik learning factory environment. Results demonstrate that our rule-based system not only achieves high accuracy in identifying waste types but also significantly reduces the time and costs associated with waste detection processes. This research represents a progressive step towards the automation of waste detection in industrial settings, promising substantial improvements in efficiency and reduction in operational costs.