基于高斯混合模型和帧间梯度信息的运动目标视频分割算法

Video Segmentation Algorithm Research of Moving Objects Based on MOGs and Interframe Gradient Information

视频序列的运动分割是运动分析和运动跟踪的基础,本研究基于高斯混合模型和帧间梯度信息提出了一种新的运动目标分割算法。在利用亮度信息对背景建立自适应高斯混合模型的基础上,进行前景的粗分割;针对由于视频信号的亮度和色彩分量随光照突变有较大的改变,导致大片背景的高斯模型产生错误匹配,误判为前景的问题,为了提高高斯模型分割算法的鲁棒性,结合结构梯度互相关函数对分割结果进一步校正,使之能适应剧烈的光照变化;最后,利用数学形态学进行后处理,消除影子和孤立的噪声点。通过不同场景的运动分割实验,表明该算法在复杂背景和剧烈光照变化条件下具有较强的鲁棒性和较高的分割精度。

Moving segmentation of Video sequences is the foundation of motion analysis and motion tracking.In this paper,a novel segmentation algorithm is proposed which is based on mixture of Gaussians（MOGs） and interframe gradient information.Firstly,an adaptive MOGs is established using luminance information of each pixel,and then a rough foreground segmentation is obtained.Secondly,luminance and chroma of each pixel are varying in a big scale,which result in illuminance changing abruptly and cause mismatch between luminance and MOGs of a pixel.And then a vast of background pixels are regarded as foreground information mistakenly.To adapt illuminance variation suddenly,an improved method combining structure gradient cross-correlation function is adopted to correct the initial segmentation.Finally,morphological methods are used to remove shadows and isolated noise pixels.Extensive experiments are performed with various video sequences,which prove that this method is robust and of high segmentation accuracy.