A Machine Vision System for Manual Assembly Line Monitoring

Customers are asking more and more customized products and expect to receive them in really short times. That is only reachable if there is horizontal and vertical integration, together with high information availability and transparency inside a company. When the production is not fully automatized, i.e. in those companies where the assembly or the production still relies on manual work of people, the monitoring of the line production, in terms of number of pieces produced, may be tricky due to the inevitable variability that operators add to the process, thus making essential the creation of smart systems able to deal with such complex environments and autonomously monitor them. Computer vision systems can be customized and very smart in various contexts, if properly modeled. In this paper we are going to describe the developed Machine Vision algorithm that, building upon BLOB (Binary Large Objects) analysis, is able to detect and count objects produced in a complex industrial context of manual assembly. The developed algorithm would then be compared to a more robust one taken as reference point for a comparison. The comparison between our algorithm and the Machine Learning-based Detector aims at showing the comparable Accuracy, Specificity and Sensitivity of our method, together with its higher versatility and processing speed, thus making it applicable in the plant-wide real-time monitoring of the manual assembly lines.