Based On The Compressed Sensing Arm Embedded Video Monitoring System

At present,video surveillance is widely used in traffic control,citysecurity,industury,national defense and so many fields for its intuition,convenience and abundant information.The traditional video surveillancecaptures video data base on Nyquist sampling theory and then compress videoimage throught algorithm of MPEG,H.26X which however brings too muchpressure for video compressing,storage,conveying due to sampling enormousvideo information and complex video compression algorithm. On the otherhand,the tranditional video surveillance has some defects such as enormoussize,complex composition,expensive price and bad real-time. the above facts,with the deep research on CS(compress sensing) theory,Embedded architectureand Linux operation system,this thesis proposes an ARM embedded videosurveillance by integrating CS video enconding optimization algorithm andembedded technology. The main work and achievement are as follow:1) A residual distributed compressive video sensing algorithm whichfeatures low-complexity encoding is proposed by combing the residual codingand CS theory based on CS and DVC theory.2) The hardware and software platform is designed based on ARMEmbedded video surveillance system.3) Some key techniques of video surveillance such as:USB camera imagecapture based on CS and the design of USB driver modules,DCVS algorithmoptimization and embed,save local video surveillance.