"

Cookies ussage consent

I agree.

Our site saves small pieces of text information (cookies) on your device in order to deliver better content and for statistical purposes. You can disable the usage of cookies by changing the settings of your browser. By browsing our site without changing the browser settings you grant us permission to store that information on your device.

LED condition monitoring system using U-Net and Luminance Flux Computing Model based on IR thermal images

M. S. KALYANA SUNDARAM1, J. GNANAVADIVEL2,* , K. S. KRISHNA VENI2

Affiliation

  1. Department of EEE, AAA College of Engineering and Technology, Tamilnadu, India
  2. Department of EEE, Mepco Schenk Engineering College, Tamilnadu, India

Abstract

This article introduces an innovative non-invasive method for health status monitoring of industrial LED lighting systems, addressing the need for reliable and efficient maintenance solutions. A combined approach utilizing a U-Net Convolutional Neural Network and a Luminance Flux Computing Model (LFCM) to identify faults in individual LEDs based on computed luminance flux values is proposed. The preprocessing unit designed in this article achieves a Peak Signal-to-Noise Ratio (PSNR) of 40.96 and precise segmentation of LED components using U-Net, achieving an accuracy of 95% and an Intersection over Union (IoU) of 90%. The proposed system effectively estimates the depreciation rate of each LED in a LED panel, providing critical insights into their health and operational efficiency. Performance evaluations reveal the effectiveness of the system and the results are compared with other deep learning techniques such as Fully Convolutional Networks (FCN), Mask R-CNN, SegNet, DeepLabv3+ and PSPNet, highlighting its potential for enhancing the longevity and reliability of industrial LED systems.

Keywords

LED health status, Non-invasive method, U-Net, Luminance Flux Computing Model, Thermal imaging, Depreciation rate.

Citation

M. S. KALYANA SUNDARAM, J. GNANAVADIVEL, K. S. KRISHNA VENI, LED condition monitoring system using U-Net and Luminance Flux Computing Model based on IR thermal images, Optoelectronics and Advanced Materials - Rapid Communications, 19, 3-4, March-April 2025, pp.179-188 (2025).

Submitted at: Oct. 14, 2024

Accepted at: April 3, 2025