摘要： 传统的视频分割通常利用低级特征提取对象。文章提出对象运动的连通性是对象的高级特征；并在累积帧差图像上，滤除尖锐噪声自适应确定域值，利用这个特性自动提取首帧运动对象。实验结果表明，该方法计算复杂度低，能够自动、快速、准确地提取运动对象，非常适于基于对象的实时视频应用。

关键词: 运动连通性；高级特征；累积帧差；自适应域值；对象提取

Abstract: Conventional video segmentations usually use low-level features to extract object. This paper originally proposes that the connectivity of object motion is an advanced feature of object and designs a novel algorithm to automatically extract the initial moving object on cumulated difference image by utilizing this advanced feature and determining adaptive threshold after eliminating acute noise. Results show that the computation complexity of this algorithm is very low and it can extract object automatically, promptly and properly. Therefore, it’s very suitable for real time video applications based on object.

Key words: Motion connectivity; Advanced feature; Cumulated difference image; Adaptive threshold; Object extraction