Abstract:

The color feature algorithm of video image is used to recognize the on and off state indicator light of circuit breaker.A typical image is divided into regions.According to the characteristics of indicator light on and off image,a 6 values gray level construction method is used to construct the image color features.Through feature collection and training of historical images in various environments,the expert knowledge base of online recognition is obtained Finally,the online video images are analyzed in real time according to the feature expert knowledge base,and the alarm is issued by comparing the preset values..Using the real-time picture data of railway traction substation in operation,the image recognition of circuit breaker on and off state is carried out in the example.The results show that the proposed method is feasible in the on-line image recognition and has certain practical value.

Key words: indicator light, image recognition, knowledge base, substation