CV-GAC模型与图割优化的运动目标检测和分割

摘要

针对视频图像序列中运动目标的分割问题,提出了一种融合测地线的水平集模型(CV-GAC)和图割优化相结合的算法.采用高斯混合模型和背景差分获取目标的运动区域,在运动区域内自动设置初始曲线的轮廓,对运动区域进行数学形态学运算,利用融合测地线的水平集模型自适应处理目标的拓扑变化,并用图割进行能量函数的最优化.实验结果表明,与其他传统方法相比,缩短了运动目标分割的时间,能够正确和快速提取运动目标的活动轮廓.%To solve the problem of moving object tracking in video image sequences,this paper presents a novel algorithm that combines the Chan and Vese Geodesic Active Contour (CV-GAC) model with graph cut optimization.Firstly,the object's active contours are obtained by using the Gaussian mixture model and background subtraction;then,the contour of the initial curve is automatically set in the moving area,and the mathematical morphological operation is carried out by using the geodesic level set model to adaptively treat the object's topology change and globally optimize the energy function with the graph cut optimization.The experimental results show that this method effectively shortens the time for the moving object segmentation,correctly and quickly extracting the active contours of moving targets.