Research On Substation Electrical Equipment Infrared Image Segmentation

Infrared imaging technology is one of the important means of substation electrical equipment monitoring.To build an intelligent substation monitoring system,a good infrared image recognition and classification technology is needed.Image segmentation technology is the key step from image processing to recognition and classification.Therefore,the infrared image segmentation technology is an important part of the intelligent substation monitoring system.It is also an important link,so it is necessary to design a good infrared image Segmentation method.In this paper,the infrared image segmentation algorithm is studied in depth,and the gray level change,image processing and image segmentation are studied.Because infrared images have the characteristics of uneven gray-level variation,lower resolution and more noise,the conventional visible light image segmentation method is difficult to divide the infrared image effectively.In order to solve this problem,two kinds of improved image segmentation methods,random walk infrared image segmentation based on Nonsubsampled Contourlet Transform(NSCT)and NSFC random walk based on fast fuzzy C-mean Image segmentation algorithm.Based on NSCT random walk infrared image segmentation algorithm,the infrared image is first decomposed by multi-dimension and multi-scale using NSCT transform to obtain the low frequency coefficient and high frequency coefficient of the infrared image,then the high frequency and low frequency coefficients of the image are processed,and then inverse transform to obtain the enhanced infrared image.Random walk image segmentation algorithm can process the image after a good segmentation,combined with NSCT can greatly improve the performance of random walk image segmentation.Because the NSCT-based random walk infrared image segmentation algorithm needs to manually give the seed point of random walk algorithm is not conducive to the implementation of intelligent systems,this paper also improved the above algorithm,and proposed a fast fuzzy C-mean based NSCT random The algorithm of walking infrared image segmentation firstly carries out the fast fuzzy C-means segmentation of the image,and obtains the seed points of the random walk algorithm by taking the central points of the initial segmentation as the starting points,and realizes the automation of the infrared image segmentation algorithm.