全局运动估计的关键在于全局运动模型的选择。结合率失真优化理论，提出一种自适应全局运动估计方法来达到编码优化的目的。该方法的主要思路是：对同一帧图像，分别使用平移运动模型、六参数运动模型和十二参数运动模型进行编码，用率失真优化算法计算3种运动模型下的拉格朗日代价函数值，拉格朗日代价函数值最小的运动模型被选为最佳的当前帧的运动模型。实验证明，该方法具有较好的鲁棒性，对不同分辨率的视频序列均有不同程度的编码增益。

The key point of the global motion estimation is how to select a global motion model. In this paper, integrating with a rate distortion optimization algorithm, an adaptive global motion estimation method to optimize the coding is proposed. The scheme of the algorithm is as follows: for the same image, using the translated motion model, affine motion model and quadratic motion model to estimate the inter frame motion and calculating their Lagrangian costs respectively. The motion model which yields least Lagrangian cost is selected as the best motion mode for current frame coding. Simulated results show that the proposed technique is robust and have better performance for various resolutions of video sequences.