Research On The Detection Method For The Hydrophobicity Of Composite Insulators Based On Image Analysis

With the increasingly serious environmental pollutions and the rapid expansion of grid scale, traditional insulators made of glass and porcelain are no longer able to meet the operating insulation requirements. Composite insulators have been widely applied in power system especially high voltage transmission lines due to their high tolerance to flashovers brought about by their superior hydrophobicity and hydrophobicity transference. As time goes by, composite insulators present different degree of aging under the influence of pollutant, electric field, ultraviolet ray and so on. The increasingly worse and worse hydrophobicity would possibly lead to flashover accident. Accordingly, it is significantly vital to test composite insulator’s hydrophobicity periodically for ensuring the safe operation of the transmission line.On the basis of related studies done by other scholars, this article study in image filtering and segmentation and classification of hydrophobic online detection of composite insulators based on image processing technology. research in interested region on the surface of composite insulator’s shedding construction, according to the inherent noise of hydrophobic image, the hydrophobic image denoising algorithm based on NSCT is put forward. Under the Nonsubsampled Contourlet framework, the low frequency noise of the image is processed with B Spline Curve to process, the high frequency noise of the image is processed with nonlinear selection filter. The results show that the algorithm not only retains useful information and also has good denoising effect.Combined with the edge and regional characteristics of targeted droplet of hydrophobic image, an image segmentation of improved GAC-CV model is proposed. Object segmentation with level set method is solved combined with different characteristics factor. The results show that the model can meet image segmentation requirement of different hydrophobic levels.This article has thoroughly analyzed the relationship between composite insulator’s hydrophobicity and flashover resistance performance. Surface electric field analysis of composite insulators based on electrical network theory is put forward. Electrical network model of composite insulators combined with the binary image after segmentation is established, considers composite insulator’s equivalent resistance and electric filed value under the different hydrophobic and contaminate degree. The results show that:under the same contaminated degree, the worse hydrophobicity, the smaller equivalent resistance, the maximum electric field values shows increase, decrease, increase, reduce variation along with the hydrophobic decline. Under the same hydrophobic degree, the maximum electric field values and the equivalent resistance are reduced along with the contaminate aggravation.Finally, the automatic classification method of composite insulators hydrophobicity based on support vector machine is put forward. Droplet features after segmentation have been extracted, including the maximum area ratio, shape factor, and box-counting dimension. These features and combination have been inputted as variables of support vector machine to analyze the classification accuracy of composite insulators’hydrophobicity. The classification results show that:when hydrophobic levels of composite insulators is divided into seven categories, considering the maximum area ratio as input parameters, the accuracy rate is the highest, reaching83.33%; when the hydrophobic level is divided into three categories, the accuracy rate reach100%considering the maximum area ratio as input parameters.