782 / 2022-03-31 22:19:50

Hydrophobic classification of composite insulators based on Faster R-CNN object detection algorithm

composite insulator,Hydrophobicity,Faster R-CNN,BP neural network,Water spray classi-fication method

Composite insulators are widely used in China's high-voltage power transmission system, but due to environmental factors and other influences, the surface of composite insulators will gradually experience insulation aging. The hydrophobicity of the surface of the composite insulator is one of the indicators that reflect its insulation status. A variety of methods for evaluating the hydrophobicity of the surface of the insulator have been proposed by various scholars. The traditional method of manually judging the hydrophobicity level of the composite insulator surface is subjective and the efficiency and accuracy are not high. However, the method of evaluating the hydrophobicity of the composite insulator based on digital image processing uses computer algorithms to reduce the workload of the evaluation process. However, the proposal of its characteristic parameters is still subjective. On this basis, this paper proposes a composite insulator hydrophobicity level detection model based on the Faster R-CNN target detection algorithm, which can intelligently complete the hydrophobicity evaluation of composite insulators, and the model allows an error of ±1 The judgment accuracy rate is close to 99%.