基于运动信息与马尔可夫彩色聚类的运动目标分割

针对视频序列图像中的运动目标分割,论文提出了将运动检测和马尔可夫彩色聚类相结合的运动目标分割算法。该算法首先利用基于统计模型的运动检测算法,通过后处理,得到运动目标的初始模板。然后,利用区域生长算法进行彩色图像的初始分割,在初始分割的基础上应用马尔可夫随机场模型进行彩色聚类,得到具有精确边缘的分割区域。最后,将运动目标的初始模板和彩色精确分割结合起来提取出具有精确边缘的运动目标。实验结果表明该算法能有效地分割和提取出视频序列中的运动目标。

Moving Object Segmentation Based on Motion Information and Color Clustering Using Markov Random Field

An algorithm extracting moving object is proposed by utilizing the motion detection and Markov color clustering in video sequences.Firstly,a rough motion template is obtained by motion detection based on statistical model and through post-processing.Secondly,after obtaining initial segmentation of the color image through region growing,clustering operation is used based on Markov random field model.Finally,the precise contour of the object is extracted by the fusion of motion template and color clustering result.The experimental results show that the proposed algorithm can effectively segment and extract moving object in video sequences.