Scene Switching Detection Based On Svc And Bit Rate Control Research

As the currently excellent video compression Standard—H.264/AVC,its outstanding compression performance and internet affinity, which is widely used in digital television, network video player, video teleconference and other fields. However, in order to adapt to different terminal.network transmission conditions and heterogeneous network, the ITU-T and ISO/IEC develop extension standard of H.264/SVC based on H.264/AVC.The encoder of H.264/SVC can produce hierarchical bit stream through temporal,SNR and spatial scalability mechanism. And you can get kinds of video quality by decoded corresponding bit stream, at the same time, H.264/SVC can improve video compression efficiency by some advanced technology, and better applied to dynamic network changes.This paper focus on that scene change existing on video sequence have impact on the course of encoding, and analysis how to detect scene change in GOP encoding unit on time and how to reduce the affection on video quality, They are as follows:(1) This paper introduced the encoder system and its key technologies of H.264/AVC, and made a detailed summary of the coding structure and three scalable mechanism about the H.264/SVC, then summarized the principle of rate control and rate distortion theory, and analyzed and compared several international classic rate control algorithm.(2) An improved algorithm based on image complexity is proposed, introduced some type of scene change in video sequences, as well contrasted several commonly scene change algorithm, combined with B frame predictive coding structure of H.264/SVC, the traditional scene change detection based on image complexity has improved, and used some experiments to prove validity and accuracy of this new algorithm.(3) Proposed a modified algorithm on the QP of the hierarchical B frames,used new algorithm to detect video sequences,for the coding unit which exist scene change, to modify the QP of B frame on the high temporal layer, and proved by experiments that this algorithm can reduce the impact of video quality which caused by scene change and did not obviously improve the bits of one GOP unit, and effectively improved the signal to noise ratio.