The Design And Implementation Of Video Streaming Distributed Storage System Based On S3C6410 And Hadoop

Compared with the traditional video surveillance system,intelligent video surveillance system have "automatic" ? "prophetic" ? "rapid response" and many other advantages.Video surveillance system usually consists of video capture front-end and the platform.It generally involves video capture and encoding,video streaming transmission,storage and backup,video intelligent analysis processes.The number of the camera in the front-end is enormous and it usually works 24 hours a day,so this requires the system to have massive storage capacity and meanwhile puts forward higher requirements on the stability,data backup,fault detection and recovery mechanisms of the system.In this paper,we design and implement a video streaming distributed storage system based on S3C6410 and Hadoop,which is combined with embedded technology,video coding and distributed storage technology.It can be used in the intelligent video surveillance.S3C6410 processor has the advantages both in control operation and digital signal processing.Its multi-format codec can support H.264 video codec,which can be used as the core processor in the front-end of the system.Hadoop is an open source distributed computing platform and its distributed file system has the advantages of high fault tolerance,high reliability and scalability and so on,which can be used to store vast amounts of video stream.The main work is as follows:(1)Firstly We use OK6410 development board as the hardware platform and implement capturing video images in the embedded Linux.Secondly,we respectively implement the H.264 video encoding based the S3C6410 multi-format codec and x264 encoding library.Finally,we choose some video streaming sequences which have different resolution,texture complexity and intensity of movement,then respectively encode them by the hardware method and software method and calculate the relevant assessment indexes to compare the video compression quality and video coding efficiency in the embedded Linux by contrast.(2)The second part is to design and implement the platform of the system.This job is mainly divided into two parts,the first one is to receive,decode and display the video streaming that transmitted from the front-end.The second one is to implement the distributed storage for video streaming based on HDFS.Meanwhile,we also have finished that multi-client can access the same front-end at the same time.Finally,we test the performance of the whole system.