Substation Infrared Image Recognition And Fault Diagnosis

With the rapid development of power industry, power capacity and voltage rating also will continue to improve, in order to ensure reliability and security of electricity supply, it is very important to monitor electrical equipment running in temperature and automatic fault diagnosis. Not only does this paper study the subject of substation infrared image processing and recognition technology , but also gives a simple fault diagnosis method.Firstly, this paper analyzes two methods on the infrared image gray pretreatment. According to the noise types and characteristics of the infrared image, this paper uses the denoising method that based on wavelet transform and mean filter, it effectively suppresses the image noise signal. Secondly, it studies many segmentation methods, and proposes the method that based on morphology combined with the Otsu and Krisch operator segmentation method .The experimental results show that this method can be applied to substation image segmentation. Then according to the characteristics of substation image, this paper uses Hu invariant moments to extract feature and propose improvements to solve the problems of the large difference in the values of magnitude about Hu moments components and discrete two imensional image doesn't meet scale invariant contraction.Experimental results show that the improved Hu moments can effectively solve the problem of the two aspects. Using the improved Hu moment extracts the features of substation equipment, establishes IR database, using the nearest neighbor classifier to identify it, and achieves satisfactory recognition rate. Finally, on the basis of the results of recognition, this paper analyzes the infrared diagnostic methods and fault reasons of thermal power equipment and its classification, and combines the temperature information database of infrared image to achieve a automatic fault diagnosis system about the infrared images that based on the relative temperature method.