Jun Chen, Wang Luo, Yunhe Hao, Huarong Xu, Jian Wu, Xiaoming Ju
{"title":"Using Convolution Neural Networks to Build a LightWeight Anomalies Detection Model","authors":"Jun Chen, Wang Luo, Yunhe Hao, Huarong Xu, Jian Wu, Xiaoming Ju","doi":"10.1109/AUTEEE52864.2021.9668683","DOIUrl":null,"url":null,"abstract":"With the development of the State Grid, the number of transmission equipment and the cost of later maintenance increases. Currently, the maintenance of power transmission equipment usually requires the dispatch of drones to take patrol photos and then detect the photos. The detection process usually requires high cost in manpower and time. Applying machine learning to the photo detection can overcome these limitations in some way. And it’s necessary to propose a lightweight model and computational method due to the limited computational resources of end devices. Therefore, this paper proposes a lightweight model, which named Lightweight convolution neural network (Light_CNN), based on deep neural network to detect anomalies in equipment pictures. The experiment on our self-constructed datasets shows that the model outperforms state-of-the-art baselines. In addition, the number of parameters and flops of Light_CNN is much smaller than other models, which can be applied to many terminal devices that are limited by computational resources.","PeriodicalId":406050,"journal":{"name":"2021 IEEE 4th International Conference on Automation, Electronics and Electrical Engineering (AUTEEE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 IEEE 4th International Conference on Automation, Electronics and Electrical Engineering (AUTEEE)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/AUTEEE52864.2021.9668683","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
With the development of the State Grid, the number of transmission equipment and the cost of later maintenance increases. Currently, the maintenance of power transmission equipment usually requires the dispatch of drones to take patrol photos and then detect the photos. The detection process usually requires high cost in manpower and time. Applying machine learning to the photo detection can overcome these limitations in some way. And it’s necessary to propose a lightweight model and computational method due to the limited computational resources of end devices. Therefore, this paper proposes a lightweight model, which named Lightweight convolution neural network (Light_CNN), based on deep neural network to detect anomalies in equipment pictures. The experiment on our self-constructed datasets shows that the model outperforms state-of-the-art baselines. In addition, the number of parameters and flops of Light_CNN is much smaller than other models, which can be applied to many terminal devices that are limited by computational resources.