Anomaly Detection for Line Clearance in Industrial Systems

With the rapid advancement of technology, industries are increasingly seeking solutions that operate efficiently, autonomously, and reliably. Line clearance, a critical procedure designed to ensure work areas are free of residual products or contaminants, is no exception, as it demands automation to enhance productivity and reduce human error. This project explores the application of anomaly detection techniques for automating line clearance across five distinct views of a machine, leveraging two neural network architectures to evaluate their feasibility and performance in operational scenarios. The study employs advanced techniques such as masking, image registration, and data augmentation to enhance the robustness and precision of the anomaly detection models. Comparative analyses focus on assessing detection accuracy across all views and the overall system reliability in industrial workflows. This research aims to establish a robust framework for deploying automated anomaly detection systems, contributing to safer, more efficient manufacturing processes. Through a comprehensive evaluation, the study provides insights into the applicability and comparative strengths of the two networks for automating line clearance and highlights their potential for broader industrial adoption.