Based On Infrared Thermal Imaging Device State Recognition Technology

The Equipment Status Recognition is the fundamental issue of the equipment fault diagnosis. With the Infrared Thermography Technology's continuous improvements, the Equipment Status Recognition based on infrared image has been attracting people's keen interests since its birth.Considering the influences made by atmospheric and environmental temperature during the infrared thermography detection, this research use the variance ratio instead of the original temperature value to acquire the image which reflects the status of equipment. The result shows that the distribution of the temperature's variance ratio on the surface of the equipment is related to its physical nature. Therefore, the distribution images of the temperature's variance ratio could be the hard evidence to judge the status of equipment.To enhance the efficiency of the equipment status recognition, this research proposes an approach of demand-oriented feature selection and using Euler numbers, edge points and barycentric coordinates as the eigenvalues of status recognition. To solve the weak target edge detection problem in specific areas of low quality infrared images, this study involves Morphological Reconstruction into the edge detection algorithm.This study designs an equipment status recognition system by using binary tree multi-class SVM as classification tool, based on theory and SVM learning. The article studies to use the infrared temperature images and the distribution images of temperature's variance ratio as the evidence to judge equipment status respectively.The results show that the equipment status identification system based on infrared thermal image could improve the efficiency of the equipment status recognition.