【摘要】 为了提高运动目标分割质量,提出一种基于时空信息的运动目标分割算法。首先利用回声状态网络仿射参数进行估计实现运动目标背景补偿,并采用非下采样轮廓波变换对视频图像进行消噪处理,然后采用标记分水岭算法对分割区域进行标记,并融合自适应光流法对运动目标进行分割,最后采用仿真实验测试算法性能。仿真实验表明,本文算法大幅度减少了运动目标分割误差,提高了分割精度,具有较强的鲁棒性。更多还原

【Abstract】 In order to improve segmentation quality of moving objects, a moving object segmentation algorithmbased on spatial-temporal information. First, support vector machine is estimated for moving target background com-pensation of affine parameters, and non-sub-sampled contourlet transform is used to eliminate background noise, andthen the marker based watershed algorithm is used to mark the partition of region, and optical flow method is used tosegment motion object, finally the performance is tested by simulation experiments. The results show that the proposedalgorithm not only greatly reduces the moving object segmentation error, improve the segmentation accuracy and haswell robust.更多还原