| With the development of computer network technology and image coding technology,daily video data generated by video surveillance systems shows a explosive growth.In this situation,massive data processing,ef ficient storage and management has become a focus of research in academ ia and industry at home and abroad.Cloud storage technology uses virtualization technology can im prove the stor age efficiency of the system.By networ k equipment from different vendors hardware storage devices conne cted together to construct a distributed storage cluster,centralized management through software,not only shared network storage resources,and multiple devices can be coo rdinated together to calculate.Therefore,cloud storage technology will gradually repl ace direct-attached storage,network storag e technology and st orage area network techn ology,which will becomes mainstream technology for massive data storage.Cloud storage technology has advantage of high capacity,high reliability,ease of scalability,etc,is suitable for storing large data.In this paper,the video surveillance system is designed based on the video storage cl oud storage system to achieve real-tim e transmission and m ass storage of video im age data.T o ensure real-tim e transmission,H.264 format video stream is encapsulated by RTP/RTCP Real-time Transport Protocol in the system.In order to achieve m ass storage of vide o data,the system uses cloud storage technology.This paper f ocuses on the overall architecture of the Hadoop Distributed File system(Hadoop Distributed File System,HDFS),using HDFS distributed storage to realize centralized management and distributed storage of video files.HDFS distributed cloud storage system has high fault tolerance,all files have multiple copies,in this case,it not only can m eet the needs of large-scale storage of video surveillance,and also can run on inexpensive hardware.According to the actual needs of the dynamic storage cluster expansion,greatly reducing operating costs of the system. |
PDF Full Text Request