Substation Equipment Status Recognition Based On Photos

As the substation develops intelligently,automatically,and unattended,it is very important to identify the substation equipment in real time and efficiently.Because the substation environment is more complicated and there are more types of equipment in the station,the traditional pattern recognition method is based on feature extraction.The real-time detection of substation equipment is difficult to achieve in real time.Therefore,it is very important to study an image-based method to identify devices in real time,efficiently and accurately.This paper systematically analyzes and summarizes the research results of predecessors,and proposes a method for accurate identification and location of equipment based on images.The research results of the thesis are as follows:1)Collecting image information of 7 common substation equipments under different lighting conditions and different climatic conditions,using the filtering method to denoise the image noise,and using the data enhancement technology to expand the image data set,thereby increasing the sample.Diversity.2)A method based on Faster R-CNN substation equipment identification is proposed.The method first uses the shared network structure to extract image features,processes the RPN network to generate candidate regions,and then uses Faster R-CNN to classify the candidate regions and use them simultaneously.The frame regression technology accurately locates the target and realizes the category identification and positioning of the substation equipment.Through the experimental analysis of seven common substation equipments,an average recognition accuracy of 87.5% was obtained.3)Propose a detection and state recognition of HF-based isolating switch.By constructing a classifier and combining a non-maximum suppression algorithm,the method uses the Hough transform framework to vote on the center position of the target object.Simultaneous detection and positioning of multiple switching devices in the image improves the accuracy of switch detection.The image-based substation equipment identification method proposed in this paper is of great significance for realizing unattended substation.At the same time,the detection of the status of the substation equipment can also provide certain reference value.