Research And Implementation Of Image CODEC Algorithm For Network Video Terminal

In recent years, with the rapid development of technology, IP cameras based on wireless or Internet are more and more widely applied. The core of IP Camera is video compression and coding. Video signals generate huge amounts of data after quantization, and digital video must be compressed in network transmission. Video images codec plays a very important status in digital video processing, transmission and storage.H.264 is the latest video standard proposed by VCEG(Video Coding Experts Group) of ITU-T and MPEG(Moving Picture Experts Group) of ISO/IEC. H.264 encoder can save 50 percent code rate compared to present video standards, maintaining at same image quality. Because of its good compression efficiency and network adaptation, H.264 will be applied in a wide range of area, such as video phone, digital TV broadcasting, video streaming media services and compressed video etc. On the other hand, digital signal processing technology rapidly develops. And high-performance general-purpose multimedia digital signal processor DM642, manufactured by Texas Instrument, is the ideal platform to implement H.264 encoder and decoder because of its powerful performance.This thesis systematically introduced H.264 standard which had superior performance compared with MPEG-4 and video standard AVS of our country. For integer transform, the core of transform was modified. Simulation showed that the new integer transform module improved the coding rate than. This paper presented the fast algorithm based on histogram, referencing the search results of intra-prediction techniques. In frame code mode, macro-block texture characteristics of their own determines intra-prediction decision mode, and then we can get the best macro-block coding mode, reducing the exhaustive algorithm complexity of the original standard algorithm. Combining DM642 chip structure with H.264 encoder characteristics, this paper completed algorithm-level, project-level optimization and programming mode optimization.. The experimental results showed that the optimized encoder had a good performance, met requirements of embedded real-time encoding.