Author:
Xu Hui,Zhao Xinyang,Yin Qiyun,Dou Junting,Liu Ruopeng,Wang Wengang
Abstract
<abstract>
<p>This study addressed the critical role of isolating switches in controlling circuit connections for the stable operation of the substation. Our research introduced an innovative state detection system that utilized depth information guidance, comprising a controllable pan-tilt mechanism, a depth camera, and an industrial computer. The software component employed a two-stage strategy for precise isolating switch detection. Initially, the red green blue with depth (RGB-D) saliency network identified the approximate area of the isolating switch target. Subsequently, a fully connected conditional random field was applied to extract accurate detection results. The real-time state of the isolating switch was determined based on the geometric relationship between its arms. This approach enhanced the accuracy of isolating switch detection, ensuring practical applicability in engineering scenarios. The significance of this research lies in its contribution to advancing isolating switch monitoring through depth information guidance, promoting a more robust and reliable power system. The key improvement is implementing a two-stage strategy, combining RGB-D saliency analysis and conditional random field processing, resulting in enhanced accuracy in isolating switch detection. As validated through extensive experiments, the proposed system's successful application in practical engineering underscores its effectiveness in meeting the accuracy requirements for isolating switch detection and state detection. This innovation holds promise for broader applications in power systems, showcasing its potential to elevate the reliability and efficiency of electrical networks. Code of the proposed system is available at: <ext-link ext-link-type="uri" xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://github.com/miaomiao0909/Isolating-Switch-Detection/tree/master">https://github.com/miaomiao0909/Isolating-Switch-Detection/tree/master</ext-link>.</p>
</abstract>
Publisher
American Institute of Mathematical Sciences (AIMS)
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献