Artificial Intelligence of Things Infrastructure for Quality Control in Cast Manufacturing Environments Shedding Light on Industry Changes

Abstract

The transition from Industry 4.0 to 5.0 raises concerns about integrating advanced quality control measures by replacing humans. The biggest challenge of this transition is infrastructure compatibility. This paper proposes a remote collaboration solution via the Internet of Things (IoT) infrastructure. The study identifies challenges in implementing such strategies and highlights the importance of AI–human collaboration, aligning with Industry 5.0 concepts. This research integrates data from multiple visual sensors (cameras) and devices into an IoT framework to create a monitoring system. This system’s application focuses on ensuring cast quality control standards. The proposed artificial AI method provides compatibility for the entire infrastructure. The Nonconformity Indicator Algorithm (NIA) was designed for defect detection. NIA, developed using Azure Custom Vision Service, identified and classified manufactured product defects based on image analysis with an Accuracy of 98.18%, Precision of 98.44%, Recall of 96.56%, and F₁-Score of 97.50%. Furthermore, an IoT-based monitoring system was designed that employs real-time sensor fusion techniques for quality control in cast manufacturing environments. The system integrates data from multiple devices, including visual sensors like the ESP32-CAM, within an IoT framework powered by Azure IoT Hub and Azure Custom Vision Service. This infrastructure enables the compatibility of devices by facilitating communication via an Azure Event Grid Trigger integrated into an Azure Function through Azure IoT Hub.