This paper presents a high-speed inspection technique based on motion-induced eddy current testing (MIECT) for detecting surface railway defects, developed in collaboration with the Maintenance Technology Center (MTC), KMUTT.
@article{brienza2026development,title={Development of high-speed inspection technique based on motion-induced eddy current testing for surface railway defects},author={Brienza, N. and Khunsombutcharoen, N. and Poopat, B. and Janya-anurak, C. and Jirarunsatian, C. and Mutthanu, R. and Aung, Htet Myat and Jomdecha, C.},journal={Nondestructive Testing and Evaluation},pages={1--22},year={2026},publisher={Taylor \& Francis},doi={10.1080/10589759.2026.2638558},url={https://www.tandfonline.com/doi/abs/10.1080/10589759.2026.2638558},}
This paper presents an optimized 3D Convolutional Neural Network (CNN) for real-time classification of natural gas leaks using infrared imaging. The proposed model achieves a high accuracy of 99.75% through the application of advanced preprocessing techniques including Running Average, Gaussian Mixture Models (GMM), Moving Average, and Background Subtraction. To enable efficient deployment on edge devices, the model was optimized using smaller image sizes (Half Size and Quarter Size), maintaining above 99% accuracy while significantly reducing computational requirements. The system demonstrates robust performance for industrial methane gas leak detection applications, providing a practical solution for real-time monitoring and safety applications.
@inproceedings{11192973,author={Aung, Htet Myat and Wongsa, Sarawan},booktitle={2025 International Conference on Information and Communication Technology (ICoICT)},title={Optimised 3D-CNN for Real-Time Infrared Natural Gas Leak Classification: Balancing Accuracy and Computational Cost},year={2025},volume={},number={},pages={1-6},keywords={Performance evaluation;Methane;Accuracy;Image resolution;Heuristic algorithms;Infrared imaging;Real-time systems;Computational efficiency;Leak detection;Optimization;Infrared Imaging;Methane Emissions Detection;Running Average Background Subtraction;Deep Learning Optimisation;Computational Efficiency},doi={10.1109/ICoICT66265.2025.11192973},url={https://ieeexplore.ieee.org/document/11192973},location={Bandung, Indonesia (Hybrid)},redirect={https://ieeexplore.ieee.org/document/11192973}}
ICA-SP
Natural Gas Methane Leak Detection Using an Infrared Camera
Htet Myat Aung
In 16th Instrumentation, Control and Automation Senior Project Conference (ICA-SP Con 2025), Apr 2025
This research presents a comprehensive approach to detecting natural gas methane leaks using infrared camera technology. The system leverages thermal imaging capabilities to identify methane gas leaks in real-time, providing a non-contact detection method suitable for industrial safety applications. The research demonstrates the effectiveness of infrared imaging combined with advanced image processing techniques for reliable gas leak detection.
@inproceedings{aung2025methane,title={Natural Gas Methane Leak Detection Using an Infrared Camera},author={Aung, Htet Myat},booktitle={16th Instrumentation, Control and Automation Senior Project Conference (ICA-SP Con 2025)},year={2025},month=apr,}
In 2025 International Conference on Information and Communication Technology (ICoICT), Bandung, Indonesia (Hybrid), 2025