State Recognition of Isolating Switch in Traction Substation Based on Dual Network
Kuan Feng, Wei Quan, Xuemin Lu · 2022 7th International Conference on Intelligent Computing and Signal Processing (ICSP) · 2022
Aiming at the problem that it is difficult to locate and identify the isolating switch in complex scenes, this paper studies a method of isolating switch state recognition based on dual network, which is divided into C-R2CNN isolating switch detection and SqueezeNet state recognition. Firstly, C-R2CNN is a rotating box detection network based on R2CNN, in which the channel attention module is utilized to effectively eliminate noise interference and extract features, improve the detection accuracy of the model, and realize the precise positioning of the isolating switch. Secondly, the isolating switch detection network is used to locate the isolating switch sub-region in the original image, then the target sub-region is cut out and sent to SqueezeNet network for state recognition, thus realizing the state recognition of the isolation switch. Experiments show that the method proposed in this paper can effectively identify the state of isolating switch.