| Substation and transmission lines is the important facilities of the power system,as the core component part of the entire power grid operations, the safety andreliability of their operations is directly related to the stability of the entire powersystem. With the construction of smart grid, power system also increases demand ofintelligent and automation; the traditional video surveillance system can not meet theneeds of the smart grid. Currently, traditional transformer substation and videosurveillance system of electric transmission line technology meet some difficulties,so it is necessary to run long-term monitoring by human factors, leading to thelimited effect of the online monitoring system and intelligent monitoring functioncan not be realized. It seriously hampered the construction of modern intelligentelectric net. Therefore, it is good idea to use computer vision and artificialintelligence recognition technology in intelligent electric net monitoring system tomake intelligent online monitoring of power system a truth. Then the electric net canoperate safely and reliably.This paper focuses on the research of computer vision technology and artificialintelligence recognition technology in intelligent video surveillance system. Theseare: Do researches on imaging preprocessing methods, mature image preprocessingtechniques according to the characteristics of the power video image system. Thenselect the applicable pre-processing algorithm as the basis intelligent videoprocessing algorithms of this paper; According to the state characteristics of indoorswitchgear substation the article proposed a intelligent algorithm to identify the stateof the switchgear. It completes image segmentation through image preprocessingtechniques and then completes switchgear state recognition taking advantage of theimproved random Hough circlet detection algorithm combined with the technologyof pattern recognition;This paper do researches on outdoor substation circuit breakersand disconnect performance characteristics of the various states and establish amathematical model, use the SIFT algorithm to extract and match equipmentparameters and then position it in a complex environment. At last it uses the Hough transform and K-NN algorithm to intelligently recognize of the state of the substationswitchgear;Do researches on the state characteristics of the transmission lines,usedigital image processing and the Hough transform, extract and recognize electricwire and further more detect and identify overlying ice, snow, and hanging pendulumof the transmission lines. It also can be accurately determine the thickness of theoverlying ice and vertical amplitude of snow line placed.In this paper, a new type of intelligent video processing algorithms arepresented and it promote the building of a strong and smart electric net, in the sametime provide technical support for real-time, automated and intelligent electric net. |
PDF Full Text Request