一种基于彩色图像分割的立体匹配算法

A Stereo Matching Algorithm Based on Color Image Segmentation

提出了一种新的立体匹配算法。首先采用均值漂移算法将参考图像根据彩色信息快速聚类成不同区域;然后根据变窗口算法计算的匹配点,把分割区域分为可靠区域和不可靠区域,计算可靠区域的模板参数;最后在全局评价函数中加入遮挡约束,通过贪婪算法把可靠区域的模板分配给不可靠区域使得全局评价函数取到最小值。实验结果表明,该算法可较好地处理大的低纹理区域和遮挡像素,匹配精度高,克服了基于局部算法中存在的边界模糊。

This paper presents a new stereo matching algorithm. First, regions of homogeneous color are located by applying the meanshift color image segmentation method. Then according to the algorithm matching points based on variable windows sizes matching method, segments are divided into reliable and unreliable segments, the planes of reliable segments are computed. Finally, the global evaluation function by adding occluded constraints, a greedy strategy is used to minimize the global energy. The experiments results show our algorithm has better performance for processing the large low texture areas and the occluded pixels, a higher matching accuracy than, can overcome the shortcoming of local algorithms.