基于改进型C-1BT变换的快速全局运动估计

Fast global motion estimation based on improved C-1BT transformation

针对存在单一运动目标视频序列的全局运动估计问题,提出一种结合C-1BT变换和自适应十字搜索法(ARPS)的快速块匹配运动估计方法。采用简单的预处理,提高灰度突然变化和存在局部运动目标时的全局运动估计的鲁棒性;简化传统C-1BT变换算法中的核函数,降低计算负担;考虑到实际视频中存在大量的零运动矢量(ZMV)区域和视频的空间相关性,改进原有的ARPS搜索法。MATLAB仿真结果证明,提出的算法在保证精度的同时极大地减少了搜索点数。

Considering the video sequences with a single moving object,this paper proposed a new fast block matching motion estimation approach,which combining C-1BT transformation and the ARPS approach.The proposed pre-processing step was simple,and could increase the robustness of the global motion estimation in the intensity suddenly changed and existence of local motion of objects.Then,it used a simplified C-1BT transformation kernel function to reduce the computational burden.In the search process,improved ARPS algorithm not only used spatial correlation,but also used a lot of early termination of the search methods.MATLAB simulation results show that the proposed approach can greatly reduce the search points while ensuring accuracy