Skip to Content
  • Proceeding Paper
  • Open Access

13 November 2026

Intelligent Conveyor Belt Inspection for Early Detection of Overheated Idlers and Visual Defects †

,
and
Innovative Systems Engineering Laboratory, National School of Applied Sciences of Tetouan, Abdelmalek Essaadi University, Tetouan 93000, Morocco
*
Author to whom correspondence should be addressed.
†
Presented at the 13th International Electronic Conference on Sensors and Applications, 18–20 November 2026; Available online: https://sciforum.net/event/ECSA-13.
Eng. Proc.2026, 163(1), 109;https://doi.org/10.3390/engproc2026163109 
(registering DOI)
This article belongs to the Proceedings of The 13th International Electronic Conference on Sensors and Applications

Abstract

Conveyors play a critical role in industrial settings and are widely used in many domains like mining, manufacturing, and material industry transport. However, unexpected failures can result in large losses for companies due to prolonged downtime in production lines, safety hazards, and production losses, particularly in unsecure locations like underground mining projects. Here, we propose an intelligent conveyor inspection system that combines infrared thermography and deep learning with computer vision for monitoring purposes. The proposed pipeline comprises four main components: first, a data acquisition system that uses both thermal and RGB cameras; second, thermal analysis to detect abnormal and overheated idlers; third, visual inspection, comparing a convolutional neural network (CNN) against a baseline Support Vector Machine (SVM) to identify belt defects and anomalies such cracks, tears, defected material; fourth, a monitoring platform that stores inspected data in a centralized database and visualizes operational insights via Power BI dashboards. Integrating thermal and visual information, the system provides a comprehensive assessment of conveyor health and enables real-time detection alerts. The experimental results obtained from a dataset of 10,307 annotated images demonstrate that the CNN achieves an accuracy of 98.21% and an AUC of 0.99 during image processing for defect detection, compared with the accuracy of 97.35% and AUC of 0.97 obtained with the SVM. Unlike existing conveyor inspection solutions that rely on single thermal monitoring or visual inspection independently, the proposed framework combines thermal diagnostics, CNN-based defect detection, centralized monitoring, real-time alerting, and a business intelligence dashboard into a unified inspection architecture. This solution provides an effective support tool that can be used to schedule earlier maintenance and that contributes to the development of improved predictive maintenance systems for industrial conveyor operations.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.