The Hydrophobicity Of The Composite Insulators Detecting Algorithm Based On The Digital Image Processing Technology

With the development of high voltage power transmission, the mechanical and electric loads stress at power transmission insulator becomes heavier than ever before. The intrinsic disadvantage of the traditional ceramic and glass insulators are more obvious as they are used in the heavy atmospheric contaminant environment.Since composite insulator can overcome the shortcomings of ceramic or glass insulator, consequentially in extra and ultra high voltage field the composite insulator has advantage.However some of difficult technical problems generated during the application of composite insulator such as the hydrophobicity on-line detection.In the present paper the hydrophobicity on-line detection method was investigated.Firstly, the development of the manufacture and research of insulator are reviewed. Compared with the traditional insulator, the advantage of composite insulator is analyzed.The hydrophobicity detection methods were concluded.The hydrophobicity mechanism is also analyzed in the present paper. The analyses show that the classical hydrophobicity directing methods such as entropy and the shape factor methods are all has some disadvantages that need to be improved.Through theoretical analyses,the hydrophobicity directing function method was presented that can be used to improve the detection accuracy.By the analyses of the previous algorithms, it can be concluded that the traditional detecting algorithm based on the digital image processing technology has some disadvantages that cannot obtain accurate detection. In order to achieve image segmentation, novel mathematical algorithm should be adopted.Employing the local histogram equalization algorithm based on homomorphic filtering can improve the display quality of the hydrophobic image of composite insulator. Initially the high-frequency noise can be removed, and then most of the remainder image noises could be filtered by the adaptive median filter algorithm.Adopting the genetic segmentation which is based on maximum class variance method, iterative search was carried out.The optimum threshold value was obtained and accurate segmentation of hydrophobic image was realized, which can meet the need of the classification of hydrophobicity levels based on BP neural network.By the improved-shape factor method the specific data for the samples was obtained,and the shortages of the method was also pointed out. Using BP neural network the classification of hydrophobicity level of composite insulator was realized. Experiment results show that BP neural network can classify the hydrophobicity levels of the composite insulator accurately.