Recognition And Detection Technology Of Iced Insulators Under Complex Environment

Insulators which are important part of electric power system play the role of mechanical connection and electrical insulation in the transmission line.Insulators are easily prone to iced due to the erosion of ice,frost,rain and snow for the long running of the outdoors all the year round,which reduced the electrical insulation strength threating the safe and stable operation of power system.The recognition and detection technology of ice-covered insulators are vulnerable to the image background,contrast and sharpness and other factors,which are directly related to the complex weather conditions.Therefore,it is very important to the safe and stable operation of the transmission line to carry out recognition and detection technology of ice-covered insulators under complex environment,which has become a pressing problem to transmission line operation and maintenance department.First an image processing method is used combining with the image processing algorithm and a set of identification system for the ice-covered insulators under a complex environment is designed.Second the ice-covered insulator image is preprocessed,mainly including image de-noising and enhancement.A method of image processing based on V-domain Retinex wavelet transform and its inverse transformation method is proposed to improve the image clarity and contrast.Then the segmentation recognition of ice-covered insulators image is carried out,and several methods of segmentation recognition are compared.The method of combining improved genetic algorithm with improved Otsu segmentation is mainly studied.The results show that the improved segmentation is superior to other methods.Finally the icing-thickness of insulator is detected and calculated.The icing condition of target insulators is judged by using the edge detection,boundary and surface pixel detection before and after the ice and ice geometric characteristics analysis and so on.Pixel coordinate can be transformed to world coordinates by using the method of camera calibration,which contributed to calculate the ice thickness of the insulators by comparing the surface and boundary distance of insulators that are freezing or not.Finally,the images of ice-covered insulators under complex ice freezing rain environment are tested and verified by using the field environment and artificial climate laboratory.It turned out that the ice-covered insulator images under different weather conditions are made as clear as possible whose definition would be improve.Haze weather effect is most obvious,increased by 45%.It has a high precision to identify the thickness of the ice-covered insulators that under different freezing rain conditions.The maximum relative error is 4.8%,and the smallest is only 1.5%.Thickness detection algorithm has a high processing speed even in the worst case that is greater than 7.06 m/s.Synthesize the above results,the icing thickness of insulators under complicated environment could be identified accurately and reliably by using the detection method proposed in this paper.