Recognition and fault diagnosis of insulator string in aerial images.

Aiming at the insulator sheds' dropping off, study on image threshold segmentation, edge detection, insulator recognition and fault diagnosis were carried out, characteristics of insulator and the background area were summarized, a new method to recognize and diagnose fault insulator was proposed. The OTSU was applied for image segmentation, and wavelet modulus maximum method for edge detection, Hough transform was improved to detect ellipsis quickly, ellipse parameters were used to design the classification standards and accurately recognize the insulator. Finally, fault diagnosis was realized based on location information. The method was verified using aerial images. The results indicate that ellipse parameters can be calculated, functions such as edge detection, insulator recognition, and fault diagnosis can be realized, and the algorithm responds quickly and accurately. [ABSTRACT FROM AUTHOR]