机构地区: 宁波大学信息科学与工程学院电路与系统研究所
出 处: 《宁波大学学报(理工版)》 2003年第2期140-144,共5页
摘 要: 提出了基于递归高阶统计和改进分水岭算法的视频图像序列中的运动目标分割算法,该算法首先利用递归高阶统计及多帧运动信息估计出图像序列中的运动区域,而后用一种改进的分水岭算法对提取出的区域进行空域分割,进而提取出视频图像序列中运动目标,实验结果表明所提出的算法能有效地提取出运动目标。 A moving object extraction method based on recursive high-order statistics and adapted watershed algorithm is proposed for video sequence. Firstly, binary motion image is achieved via recursive high-order statistics detection method and motion information of multiple frames. Then, an adapted watershed algorithm is used to segment the motion region. Finally the moving object can be extracted from video sequence. Simulation results show that the proposed automatic object extraction algorithm is efficient.