Abstract: Aiming at the shortcoming which the computation of target bits and Quantization Parameters(QP) in the situations of high motions and scene change of the existing H.264 rate control, this paper proposes a method of frame-level target bits allocation which based on the frame complexity to improve the above-mentioned two aspects. Due to the characteristic of moving image, which the differences in difference image is obvious, it uses frame complexity to re-allocation the target bits, at the same time, uses the frame complexity to detecting scene change to adjust the quantization parameters. Experimental results show that the improved algorithm can make target bits more closer to the actual bits, increase peak signal to noise ratio, decrease fluctuation of peak signal to noise ratio.

Key words: bit rate control, bits allocation, complexity, Quantization Parameter(QP)

摘要： 针对现有H.264码率控制在剧烈运动场景和场景切换时，目标比特数和量化参数计算存在的缺陷，提出一种基于帧复杂度的改进策略，根据图像序列的运动性和帧间差异显著的特点，利用帧复杂度对目标比特进行重新分配，同时利用帧的复杂度检测场景切换来调整量化参数。仿真结果表明，该方法能使目标比特数更接近实际编码的比特数，提高峰值信噪比，减少峰值信噪比的波动。

关键词: 码率控制, 比特分配, 复杂度, 量化参数

CLC Number: 

• TN919.8