中文摘要：

为解决视频图像在互联网中进行传输时，其质量易受网络丢包率、时延等因素的影响而显著降低的问题，提出了一种基于丢包率预测的视频传输纠错算法。该算法采用隐马尔可夫模型预测网络丢包率，根据丢包率的大小自适应地选择FEC或ARQ对视频图像进行纠错操作。当预测出的丢包率较高时，为避免FEC算法在丢包率较高时降低带宽利用率，采用选择性ARQ算法恢复丢失的视频数据包，并通过限制其重传次数使视频传输的实时性得到了保证；当预测出的丢包率较低时，则采用优化了RS冗余值的FEC算法进行纠错操作。在OP-NETmodeler中进行的仿真实验表明，与HARQ算法相比，使用该纠错算法，视频图像的PSNR的平均值提高了1．6dB，平均时延减少了0．24s左右。该算法不但降低了视频传输的平均时延和丢包率，而且提高了接收端视频图像的重建质量，具有复杂度低、实现简单的特点。

英文摘要：

For the compressed video bit stream transmission over Internet was prone to packet loss and delay, this paper pro- posed a predicting packet loss based an error control algorithm （PPLECA） in video transmission. It used a hidden Markov model to predict future packet loss statistics. According to the statistics, it adaptively selected the forward error correction （FEC） or automatic repeat request （ARQ） to perform error correction on the video data. If the statistics of Internet packet loss was large, then the ARQ algorithm which the number of retransmission was restricted was used to avoid reduce bandwidth utilization, otherwise the FEC which the amount of redundancy was adjusted was utilized by sender. The experiment in OPNET modeler reveals that the average PSNR of PPLECA h!gher than HARQ 1.6 dB, and average delay decreased 0.54 s. This indi- cated that PPLECA outperforms the HARQ in terms of both packet loss and delay, and provided a more stable video transmis- sion quality with low complexity and easy to implement.