Method for detection the non-soluble deposit density of insulators based on hyperspectral technology

Abstract

The rapid development of agriculture and industry in China leads to a fast increase in the non-soluble deposit density (NSDD) of insulators. At present, the detection of NSDD is concentrated in the use of optical principles, which is difficult to achieve quantitative detection. Hyperspectral technique is a new comprehensive image data technique based on imaging spectroscopy, which has the advantages of multi-band, high resolution. Therefore, a non-contact detection method for detecting NSDD based on hyperspectral technique is proposed. The hyperspectral images of the insulator were obtained and preprocessed with black-and-white correction, which were used to establish the model based on the extreme learning machine with the kernel (KELM). Finally, the detection of NSDD was realized. Consequently, this study can guide the prevention of flash and configuration of external insulation in transmission lines.