Research On The Intelligent Extraction Method Of Insulator Area Of Infrared Image

Insulator is important power supply equipment cuting off from high voltage current-carrying conductor,supporting structures and the the eart.Insulators are widely used in each pole,each tower and transformer and other power supply equipment.Due to the impact of the long atmospheric pollution,hot and cold changes and rain erosion,the running insulation may occur deterioration.Once appeared deterioration insulator in the insulator strings,it will reduce the effective insulator strings creepage distance and increase the probability of flashover insulator strings,the direct result is that insulator strings broken,wires and other chains landing accident,a serious threat to the security and stability of the power system run.So insulator state detection is an onerous task.Compared with the traditional artificial climbing pole line detection method,patroling the line taken by infrared thermal imagerand using image processing technology to automatically detect fault line is more efficient,it is currently the main direction of development of intelligent transmission line inspection.In view of the present low insulator,zero test method of missing of rate is high,complex operation and so on,this paper proposes a new intelligent extraction methods of insulator disk and steel cap to obtain the image texture and temperature of the disk and steel cap area and other information to achieve insulator infrared image automatically and batch processing,laid a solid foundation for insulator detecting state.Aiming at the characteristics of infrared image complex background,this paper proposes an insulator string automatic location method based on gray stretching and Hough transform.Firstly,the infrared image of insulator string is enhanced by gray level stretching based on Otsu threshold;the gray image open-operation is carried out to denoise the whole image;then gray Otsu threshold method is adopted to segment the image;after the binarization of the image,maximum connected region filtration method is carried to retrieve the region of insulator string;at last,the linear regression is used to do the image gradient correction,achieving insulator sting final automatic location.Experiment results show that the proposed segmentation can retrieve the target fast and perfectly;the proposed location method can achieve insulator string location under complicated back-grounds automatically and accurately,and it lays a foundation for the intelligent testing of the deteriorated insulator by infrared technology.Based on Jiangxi Province degrade insulator with case analysis is validated with image segmentation algorithm on the intelligence,the infrared image gray level change in the first place,the enhancement algorithm is presented in this paper to image enhancement of gray image,binarization,threshold segmentation,filtering to extract the insulator location in the region and to extract the insulator steel cap and the location of the disk,according to the steel cap and disk location to extract the corresponding temperature and gray information,it is concluded that degrade insulator fever and gray level.The results showed that:this method can infrared images from the scene of the complete substation extract the insulators' disk and steel cap area,there is a higher value in engineering.