Design and Implementation of a Real-Time Automated Optical Inspection (AOI) and Energy Management Framework for Smart Industrial and Building Automation Systems

Modern industrial manufacturing lines and commercial facility infrastructures increasingly demand tightly coupled quality assurance and energy-aware actuation. However, a persistent operational gap separates high-throughput visual quality inspection from legacy programmable logic controller (PLC) and building management system (BMS) networks. Commercial automated optical inspection (AOI) platforms frequently rely on proprietary hardware and computationally intensive deep neural networks that require dedicated graphics processing units (GPUs), creating prohibitive capital costs and nondeterministic communication latencies when interfaced with fieldbus protocols. To address this challenge, this paper presents the design and implementation of a lightweight, deterministic, real-time hybrid AOI and energy management framework tailored for smart industrial and building automation systems. The proposed three-tier architecture integrates an open-source Python and OpenCV edge computer vision pipeline—incorporating Gaussian smoothing, Contrast Limited Adaptive Histogram Equalization (CLAHE), Canny edge detection, morphological gradient filtering, and contour-based geometric segmentation—with industrial PLC and KNX building controllers communicating over Modbus TCP/IP and OPC-UA protocols. Furthermore, a real-time supervisory human-machine interface (HMI) and telemetry dashboard is developed using Streamlit and Plotly for continuous Overall Equipment Effectiveness (OEE) tracking, automated fault logging, and multi-zone power optimization. Extensive hardware-in-the-loop (HIL) evaluations demonstrate that the proposed edge framework achieves a surface and component defect detection accuracy of 98.4%, a precision of 98.1%, and a recall of 98.6%, while maintaining an average end-to-end inference latency of 34 ms on standard edge hardware without GPU acceleration. Simultaneously, closed-loop PLC-BMS interlock scheduling yields an 18.5% reduction in anomalous standby power consumption across automated production and facility zones.