提出了一种基于线段分割的图割立体匹配方法.首先,把图像的扫描线分割成不相重叠的、彩色-空间一致的线段基元.然后,在能量最小化框架下利用图割优化算法实现立体匹配,采用图割方法得到全局最优解,本算法中优化的节点不是传统的像素,而是分割得到的线段单元.图割方法是在拟合能量最优化的情况下给每个线段单元分配对应的视差来实现立体匹配的.多组测试数据上的实验结果说明了本方法的有效性.%A novel line-segments based stereo matching algorithm using graph cuts is proposed. Firstly, the reference image is divided into non梠verlapping homogeneous line segments along scan line. Secondly, the stereo matching problem is formulated as an energy minimization problem in the line segment domain instead of the traditional pixel domain. Graph cuts technique is used to speed up approximating the optimal solution, which assigns the corresponding disparity to each line segments. Experimental results on various testing data sets demonstrate the good performance of the proposed algorithm.
展开▼
