Comparative Study of Thresholding Techniques for Thermographic Diagnosis of Electrical Equipment

Abstract

This article aims to analyze the performance of different thresholding techniques to detect the hotspot of electrical equipment through Infrared thermography (IRT). Finding a suitable thresholding technique for thermal images is essential for automatic thermal condition monitoring of electrical equipment. In this paper, five different thresholding techniques (i.e., Maximum Entropy/Kapur, Minimum Error, Moments, Renyi Entropy, and Yen Thresholding) have been applied to twelve different samples of thermal images of electrical equipment to measure the overall performance. Thermal Images of electrical equipment are taken using an infrared camera from a ready-made garment factory situated in Dhaka, Bangladesh. The results show that the Moments thresholding technique has the lowest error percentage, whereas the Minimum Error thresholding technique shows the lowest accuracy.