The Research On Streaming Media Middleware And Video System Based On Internet Of Things

This thesis designed a video system based on the Internet of Things. Itadopts the AVS standard, which is developed by China independently. Thesystem consists of video sensor server, streaming middleware and videodisplay client software. Through designing each module, it realized a videosystem of real-time, clear and can be used in LAN, WAN networkingFirstly, design and implement a point to point video processing andtransmission systems, including a video sensing server and a decoding and-playing client. Wherein the video sensing server was built on TI’s DaVinciplatform TMSDM6446, Mainly using the development board’s DVSDKsuite driver to control the sensor camera to capture video; Through CCS,package and migrate AVS encoder on the DSP development board to encodeand compress video stream; Using SIP protocol and RTP protocol to transmitdata on the network. Decoding play client software supports Windows andLinux platforms. Use MFC to design Windows interface, and decode thereceived stream of compressed data by the FFMPEG library, then completedata display through Directshow; Use QT to design Linux platform interface,and decode streaming data though GAVS, then display image by SDL. Andon the two platforms local multi-screen player are designed. Video transmission system on Internet of Things typically include manyvideo sensing servers and decoding play client softwares, involving twodifferent transmission methods: direct media streaming and file generationstreaming transmission, so it requires streaming media middleware solutionfor information exchange and communication.On the Linux platform, a streaming media middleware was designedand achieved,which is located lower on the application program to manageand allocate video streaming data, which solved the communication blockingproblems caused by too much communication lines and excessive data,,making the entire system more reasonable and data transfer more smoothly.Streaming middleware use SIP protocol and RTP protocol as the datatransfer mechanism to control the scheduling of video stream data; Throughconfiguring Opensips server, NAT traversal in different networkenvironments was realized; Proposed and implemented a distributedmiddleware solutions, enabling the system to deploy more video terminalsand record video sensor data sent from the server on the network so thatusers can access historical video data anywhere, anytime.Through the entire system performance test, the system can ensuresmooth real-time D1resolution, with significantly fewer network resources.