• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

ELECTRIC POWER CONSTRUCTION ›› 2013, Vol. 34 ›› Issue (11): 121-124.doi: 10.3969/j.issn.1000-7229.2013.11.024

Previous Articles     Next Articles

Detection of Transformer Oil Leakage Based on Image Processing

DONG Baoguo   

  1. Hunan Electric Power Design Institute, Changsha 410007, China
  • Online:2013-11-01

Abstract:

In unattended substations, as power transformer serves as common and major equipment, its operation status should be monitored at all times, and corresponding alarms should be triggered when abnormal situations occur. A detection method for transformer oil leakage based on color feature was presented, in which abnormal area could be detected through difference method, and its H-S Color Histogram could be analyzed. When the tawny part of images before and after changed much, it could be determined that oil leaks abnormally. The experimental results show that this method is easy to carry out, the accurate detection of transformers oil leakage, and it can also well eliminate the influence of shadows on testing.

Key words: power transformer, color feature, transformer oil leakage, difference method, abnormal area