Research On The Techniques Of Massive Video Management Based On The Cloud Platform

With the peace of the city for nearly a decade, a large number of monitoring video storage and information extraction subjected to enormous challenges. Vast amounts of video especially high-definition video, particularly large data processing technology are needed to deal with. The valuable information extracted from video, these valuable information to all walks of life to provide services such as transportation, public security, national defense. Due to the continuous video monitoring, coupled with the emergence of high-definition monitor equipment, a day will generate terabytes and petabytes massive video data, traditional database has been unable to meet the requirements of data storage and analysis.Based on the above situations, this paper tries to use Hadoop cloud platform for massive video data storage and processing. HDFS first emphatically analyzes the Hadoop distributed storage system, key technology such as parallel computing framework graphs and YARN,starting from the mass video storage and analysis, design and develop a massive video management system based on Hadoop. Papers from distributed HDFS cluster platform structures, a distributed storage system and parallel computing engine, massive video data management at the core of the service and so on several aspects describes the overall architecture of the system, In the further study of key technology on the basis of each layer,realized the user management, cluster server management, video management, distributed video processing and other services.Secondly this paper designed the three kinds of video distributed processing strategy, the frame of distributed, single point cutting, distributed cutting strategy, and carries on the corresponding test and contrast, through the selection of the experimental results a treatment strategy as a whole system of the most efficient model. This paper through to the massive video management system based on cloud platforms of design and practice, from the performance, the computing performance and system stability, and many other aspects than the traditional video management system has a very big promotion and improvement.