摘要
针对现有立体匹配算法复杂、耗时较长的特点,提出了一种基于加权匹配代价和色彩分割的立体匹配算法。算法引入一种基于梯度差算子与绝对差算子加权的匹配代价计算初始视差。根据基于色彩分割的视差假设,采用Mean-Shift算子对图像进行分割,用分割后的图像作为模板优化初始视差。实验结果表明,该算法提高匹配精度和速度,具有良好的实用价值。
The existing stereo matching algorithms are complicated and time-consuming, a novel local method based on weighted matching cost and color segmentation is presented. Firstly a matching cost weighted by absolute differences of gradient and sum of squared difference is applied to calculate the initial disparity of each pixel.Secondly according to the assumption of disparity based on color segmentation, the image segmented by Mean-Shift algorithm is used as a model to optimize the initial disparity.Experimental results show that the proposed algorithm can not only improve the matching accuracy but also increase the matching speed, and the method is of great practical value.
出处
《无线电通信技术》
2015年第2期51-53,75,共4页
Radio Communications Technology
基金
国家自然科学基金项目(61175111)
扬州市2012年产学研合作专项(2012038-8)
关键词
立体匹配
匹配代价
色彩分割
视差
stereo matching
matching cost
color segment
disparity
