Journal of Shanghai Jiaotong University ›› 2013, Vol. 47 ›› Issue (12): 1818-1822.

• Automation Technique, Computer Technology • Previous Articles     Next Articles

Contour Extraction and Fault Detection of Insulator Strings in Aerial Images

FANG Ting,DONG Chong,HU Xingliu,WANG Yan
  

  1. (1. School of Electrical and Information Engineering, Anhui University of Technology, Ma’anshan 243000, Anhui, China)
  • Received:2013-05-13 Online:2013-12-28 Published:2013-12-28

Abstract:

Based on the fact that insulator bursting and missing are threats to the safety of electric power, this paper proposed an algorithm to detect fault by calculating the number of insulators. Because of the fixed number of insulators, an image processing method was adopted to detect the fault of insulators by calculating the number of insulators, which can be applied to intelligence patrol system. Otsu and morphological filtering were used to preprocess the image, the immune genetic snake algorithm was adopted to optimize insulator concavity regions to get a complete contour of the insulator strings, and the ellipse fitting and connected region algorithm were used to calculate the number of string of insulators. The proposed algorithm can extract the initial contour of insulator strings effectively and calculate the number of insulators, establishing a good foundation for further study of insulator fault detection.
 

Key words: insulator contour extraction, genetic immune algorithm, Snake model, fault detection

CLC Number: