Automatic inspection of outdoor insulators using image processing and intelligent techniques

One of the methods used to inspect and assist nonceramic insulator surface hydrophobicity is the STRI hydrophobicity classification guide, which is currently manually employed in the field. In this paper, a system for the automation of the STRI hydrophobicity classification system has been developed. The proposed system can be divided into two main parts: hardware and software implementation. The hardware consists of a digital camera, sprayer and a microcontroller. The main task of the hardware is to capture an image of the silicone rubber coated surface after spraying it with water. Through a communication channel, the image will be transferred to a computer for further analysis. The software main task is to process the acquired image of the insulator surface and eventually classify its hydrophobicity into one of three levels. The image processing involved here is extracting texture features using the co-occurrence matrix of the edges in the image. The extracted features are classified using the polynomial classifier. The performance of the proposed system is assessed on a dataset of more than 350 images where a classification rate of more than 97% is achieved.