Method for Identifying, Locating and Diagnosing the Status of Indicator Lights in Panel and Cabinet Based on Multi-Feature

Yaxin Yang, Kaicheng Guo

Abstract: In order to solve the problem that the traditional substation is not connected with the integrated automation system, which leads to the problem of false detection and missing detection of equipment fault, this paper proposes a method for identifying and locating and diagnosing the status of indicator lights in panel and cabinet based on multi-feature. Firstly, the color histogram algorithm was used to extract the color features of red, yellow and green indicator lights when they were on or off; Secondly, based on the color image threshold segmentation algorithm in the video stream, the real-time contour of a single or multiple fault device indicator area was obtained, and the centroid coordinate position of the area was calculated to realize the fault device indicator location and fault type judgment; Finally, the video stream analysis and statistics were used to obtain the time sequence characteristics of indicator light on and off. The video images of red, yellow and green indicator lights flashing at 0.5Hz, 2Hz and 0.5-2hz frequency are simulated. The detection results show that the detection accuracy is higher than 85% except for the low recognition rate (81.8%) of continuous change of indicator light flashing frequency. This method can accurately identify and locate the indicator light of the faulty equipment in the intelligent patrol panel of the main control room of the substation, and judge the type and emergency degree of the faulty equipment, so as to provide auxiliary technical support for the dispatching automation system of the traditional substation.

Keywords: Threshold Segmentation, Regional Characteristics, Time Series Characteristics, Video Image