摘要
网络视频解码的基础是码流的管理和视频流的解析。所设计的基于局域网的视频解码算法采用了码流的三级缓冲管理机制 :由接收缓冲区负责存储从服务器端接收到的码流 ,工作缓冲区直接向视频解码提供所需数据 ,中间缓冲区负责以上 2个缓冲区的中转。码流的解析首先从复合流中提取视频流 ,然后从视频流中读取一定位数的数据进行解码。为了保证视频解码过程的随机开始 ,在解码开始前 ,要求服务器先发送有关信息以便得到解码所需的必要参数 ,再接收码流并从找到的第 1帧I图像开始解码。为了正确解出紧跟在第 1帧I图像后面的B图像 ,系统构造了前向预测图像数据。
The management of the bitstream and the parsing of video data are the basis of the video decoder. Video decoder algorithm based on LAN adopts three kinds of buffers were offered to store the bitstream. The receiving buffer stores the bitstream come from the server. The working buffer directly gives needed bits to the decoder. The middle buffer associates the above mentioned two buffers. The bitstream parsing consists of two steps. Firstly the video data must be extracted from the compound bitstream, then required bits are provided for the decoder. Since the decoding was asked to begin at any time, the client requests the server to transfer the information to fill the indispensable parameters before receiving the bitstream, furthermore the decoder ignores all the bitstream until the first I frame comes. In order to decode the B frame following the first I frame, the algorithm creates a previous prediction picture. The Bayer ordered dithering was used to show the sequential pictures nicely. The algorithm have been used to video monitor system of bank, and have good stability. Offers a satisfied resolution for the similar applications.
出处
《中国农业大学学报》
CAS
CSCD
北大核心
2004年第2期73-77,共5页
Journal of China Agricultural University
