Intra Prediction And Transform Algorithms Research Based On H.264 Standard

The digital image processing techniques are applied to the videophone, video conference, supervisory control system, intelligence traffic monitoring, target tracking , robot navigation, etc. broadly area contain civil, commercial and industry field with the fact that digital multimedia ceaselessness developing .Its basal principle is that the using video compressing technology to compress the analog signals which are different by the video collection cell, then transmit to the sink to do some operations such as decoding, playing and storing etc. with the computer networks. Now the developmental trend is to research the method to digitalize the analog signals and study the video stream that correspond to the compressing standard, which based on the embed technology. It is easy to meet the local area network and realize the dynamic digital monitoring technology. Especially, the bandwidth bottleneck problem becomes outstanding while facing the network transmission. Therefore, we need an appropriate coding/decoding standard.H.264 standard is the latest international video coding standard. There are many advanced video compressing coding method. It has made obvious progress compared with former video coding standards. Based on the same vision apperceive quality, H.264 standard has improved coding efficiency by 50% compared to MPEG-2 standard. Moreover, it has much better network friendship. However, high coding compression needs high calculation complication, the calculation complication degree of H.264 standard is 3 times as H.263 standard. So, it needs to do some optimization on the algorithm to improve the coding efficiency in real applications.The thesis has studied the basic principle video compression technology broadly, especially a train of international video coding standards. It takes H.264 video coding international standard as the researching object. Based on deeply studied the algorithms of H.264, it did some research and analysis on the complexity and real-time ability of H.264 intra predictive modes. Then, the thesis put an efficient method forward, which uses the prediction modes of Intra₁6×16 to instruct the Intra₄×4 prediction. It could not only improve the coding efficiency, but also decrease the coding time.The thesis has done some optimition on the JM86 test model. And then it examined the model which optimized with 10 standard qcif video sequences. The final experiment has also indicated that the algorithm brought forward by the thesis has an effect to save coding time, and ensured the image quality after optimized the coding algorithm at the same time. That is to say, the decline of SNR would not exceed 0.24dB while the coding time has decrease obviously. Since the algorithm could be used to different video and received fairly good effect, it also has the extendibility.The thesis has done some research and analysis on one key technology of H.264 standard which was called integer DCT transform method at the same time. At last, it summarized the superiority of the H.264 integer DCT transform method.