探讨了提高H.264视频编码性能的场景自适应比特率控制问题。通过计算高维空间中多模式聚集的最大特征向量距离,提出了基于巴氏距离的场景变化度量,并有效采用低秩Cholesky分解计算近似巴氏距离。在此基础上,提出了自适应的GOP结构和量化参数调整策略。实验模拟了视频样本序列。由于具有差错复原和较低的计算复杂度等特点,实验结果表明该方法十分有效。 This paper discusses the scene adaptive bitrate control problem to improve the performance of H.264 video coding. By calculating the maximum eigenvector distance of multi-mode clustering in high-dimensional space, a measure of scene change based on Pahschuch distance is proposed, and the approximate Pahsvalue distance is effectively calculated by using low rank Cholesky decomposition. On this basis, an adaptive GOP structure and quantitative parameter adjustment strategy are proposed. Experiments simulate the video sequence of samples. Due to the characteristics of error recovery and low computational complexity, the experimental results show that this method is very effective.