Research On Audio And Video Compression And Transmission In Embedded Systems

The development of multimedia technology has greatly advanced the society. Image processing and media transmission are two key technologies in the domain of multimedia, and have been widely used in daily life, industry and military affairs. With the intelligentization of all kinds of electrical and mechanical productions, the embedded technology has made great progress, and embedded systems can be seen everywhere. Therefore, it will be very promising to research Image processing and media transmission based on embedded systems.In this thesis, the quality lifting technics of transform-coded images are considered. Researchs on media transmission are also included. Afterwards, task scheduling technics in real-time systems are researched. Finallyt, this thesis introduces the design and implementation of embedded real-time audio and video monitoring system.While compressing images with DCT(Discrete Cosine Transform), blocking effects will come up, which reduce the quality of images. In this thesis, a simple and efficient algorithm is designed to dispose the blocking effects without blurring the image.QoS(Quality of Service) guarantee mechanism while transmitting media data over Internet is researched. A way to adjust the bitrate of video encoder according to information fed back by RTCP is analyzed. Then, a rate control mechanism based on the end-user's device constraints is introduced, which can improve the quality of media transmission without causing network congestion.After analyzing several typical task scheduling algorithms in uni-processor real-time system, this thesis gives a multi-task scheduling algorithm based on DMS(Deadline Monotonic Scheduling) and EDF(Earliest Deadline First), which is efficient in scheduling several different types of tasks.Finally, the design and implementation of embedded real-time audio and video monitoring system are discussed. This thesis gives the top-level structure of the system and details of sub-systems in it. In the system, DCT implementation is optimized with DSP's custom operations. This thesis also introduces the synchronization mechanism between data compressing module and media transmission module while accessing shared data.