Research And Implementation Of Video Denoising In Communication Base Station Monitoring System

With the increasing of demands for network reliability and quality ofcommunication, communications put forward higher requirements for the base stationmaintenance, and the traditional manual maintenance mode can not meet the needs ofrapid development. Therefore, it is of great significance to develop a comprehensiveoutdoor base station unattended monitoring system. The quality of surveillance videonot only affects the ornamental, but also affects the information content of the videoimage, which is directly related to the accident location of the base station and theresponsibility affirmation. Thus improving the quality of monitoring video is one of theproblems to be solved.Firstly, the basic theory of video denoising is systematically studied, and Non-localmeans spatial domain filter method and its advantages compared to other traditionalmethods are discussed in detail. On this basis, taking into account the issue of temporaldomain denoising, a novel spatio-temporal combination method for video denoising isdesigned. What is more, adaptive filter strength is introduced to present a modifiedspatio-temporal combination method. Experiments show that both algorithms haveachieved good results.Secondly, for the characteristics of fixed background of the surveillance video, avideo denoising method is proposed by combing the background extraction and thefiltering of foreground. Background difference method based on noise estimation isused to separate the background and the foreground of the video images, and then theforeground is filtered using the spatio-temporal combination video denoising method.Experiments illustrate that this method works well for the surveillance video denoisingwhich has the characteristics of fixed background and greatly reduces thecomputational complexity.Thirdly, an outdoor base station unattended monitoring system is designed usingi.MX53based on a detailed analysis of the base station equipment and base stationssurroundings. The system realizes the function of video monitoring and comprehensivemonitoring of the fire, security, and power environment of the base station. Finally,videos collected by the monitoring system are denoised by the methods proposed in thisthesis, experimental results show that the proposed methods are effective. Finally, a summary of the work is made and further research suggestions areproposed in the end of the thesis.