Applying CNN to infrared thermography for preventive maintenance of electrical equipment

Abstract

This paper presents the new method to classify the condition of electrical equipment such as main circuit breaker, magnetic contactor, in term of thermal radiation effect using infrared thermography. The conventional neural network (CNN) is one of deep learning method which is widely used in pattern recognition and object detection. In this paper, the thermal image processing was used to determine the critical temperature on equipment and, the deep learning technology was applied to identify the type of equipment. Therefore, we can know the condition with the type of electrical equipment for maintenance purpose in real-time. An accuracy of our method is 91% for identifying type of equipment. This technique can be implemented to an automatic alarm annunciation system for other dangerous equipment with sensitive to thermal as well.