Aiming at the loss of rate-distortion performance in the high-efficiency video coding tree units (CTU) layer rate control scheme, this paper proposes an improved CTU-level rate control algorithm in the spatial domain. Specifically, in the CTU layer, the Scharr gradient detection operator and the Hadamard transform algorithm are used to detect the texture complexity area and the transform compression area, respectively. The scale coefficients are normalized and adaptively adjusted to construct a new adjustment factor for the gradient and sum of the absolute transformed difference of the image. In this way, the target bits of each CTU layer are reasonably adjusted and accurately allocated. Meanwhile, a more accurate model parameter update is realized in the coding structure of the CTU group. The experimental results show that compared with the dynamic adjustment control algorithm under configurations of low-delay P and B frames, the relative error of code rate of the optimized algorithm is both reduced by 0.012% compared with the dynamic adjustment control algorithm, while the rate-distortion performance is improved by 2.4% and 2.6%, respectively.

针对高效视频编码树单元（CTU）层速率控制方案中速率失真性能的损失，本文提出了一种改进的空间域CTU级速率控制算法。具体来说，在CTU层，分别使用Scharr梯度检测算子和Hadamard变换算法检测纹理复杂度区域和变换压缩区域。对比例系数进行归一化和自适应调整，为图像的绝对变换差值的梯度和总和构造一个新的调整因子。这样，每个CTU层的目标比特就得到了合理的调整和准确的分配。同时，在CTU组的编码结构中实现了更准确的模型参数更新。