The Research Of Intelligent Remote Video Monitoring System In Unmanned Substation

The development of electrical industry requires realizing the unmanned supervision in substation in deed,on the other hand,the man-made sabotage in electric network has reached the degree of record-breaking. The crimes of stealing and destroying electrical equipment cut the power supply,but also lead to huge economic losses, even threaten the safety of the public.There are a few substations equiping remote supervision systems, but they mainly rely on the manual supervision. The operators'attention is likely to scatter after watching for a long time, thus, they probably neglect some circumstantialities. Usually, there are several scenes to be watched at the same time, the effect of manual supervision get worse.Aiming at the defects in present remote supervision of unmanned substation, this paper presents a new thought introducing visual analysis into remote supervision of unmanned substation to improve the level of surveillance. After some methods being compared, a cauchy statistical model based algorithm is chose, and the definition and maintenance of the coefficients of the model have been designed.The cauchy statistical model based algorithm, warning function and the function saving the picture at the warning time are programmed with C++. Linux that has been cut down forms the operate system of the embedded remote supervision systems. On the basis of reading and understanding the source code of Gnomemeeting, this paper modify Gnomemeeting, which includes adding the function carrying out the analysis to the video and the function saving the picture at the warning time, removing the redundancies in Gnomemeeting, thus forms the application of the embedded remote supervision systems.An experiment using ordinary camera was implemented; the whole system is tested in LAN. Under many cases, the experiment results show good performance: the system can adapt the gradual and sudden change; can remove false warning introduced by the branches in the wind; can adapt windy, rainy, and snowy weather. The system shows low false warning probability.