The Design And Implementation Of Video Surveillance System Of Phtovaltaic Power Sation

Photovoltaic power stations are often built in populated and complicated areas,which brings a large number of difficulties to equipment maintenance and personnel management.The traditional photovoltaic power station video surveillance system only provides the long-distance transmission of video and image,however,people need to check several emergency situations by surveillance screen by themselves which not only increases the human cost,but can not give feedback to the staff in time when the fire or theft happened on equipments because of the limited human resource and insufficient energy.Therefore,aiming at the problem that the PV power station video monitoring system can not give the feedback about the abnormal events in time,this thesis propose an intelligent video monitoring system,which can instantly informed the relevant people when some emergencies happened on the PV power station.Firstly,the whole scheme of the video monitoring system is designed according to the function demand of the video monitoring system under the application background of PV power station.The hardware platform of the system uses the Development board based on S5PV210 microprocessor which is regard as the core to improve the development speed.In the aspect of software designing in this system,the system's software development environment is built through the transplant of system's porting and OPENCV database,which prepares the detection of video surveillance system's target moving,video acquisition and transmission.Secondly,the advantages and disadvantages of several common moving object detection algorithms will be analyzed in this thesis,and proposes a new algorithm based on five frame difference method and improved mixed Gaussian model according to he complex environment of practical application.This algorithm respectively improves the background model's updating and the foreground moving target segmentation and combining five-frame difference method to the light mutation sensitivity characteristic which can effectively improve the detection accuracy which happened in algorithm in the complex environment to the movement target.Then,the specific functions of video surveillance system should be designed and realized.Video image acquisition design is based on V4l2 drive architecture,and the Computer Vision Library OpenCV can achieve the detection of moving objects.Thecompression coding of video image and the network transmission of video image are based on H.264 format to compress and encode the video image,then the H.264 code stream is sent to the network through RTP/RTCP transmission protocol,which can realize the network transmission of video image.Finally,the experiment results show that the video monitoring system can automatically detect the moving target in the complex environment,and realize the real-time remote video monitoring,which has achieved the expected design goal and has high practical application value.