采用色彩相似性约束的立体匹配

Stereo matching algorithm using constraint of color similarity

在立体匹配算法中引入色彩相似性约束条件。首先加入权值模板来对用于灰度空间匹配的SSD算法进行改进,在抑制噪声的同时,可以最大程度的保持景物的细节信息,从而有效的解决SSD算法对于噪声比较敏感的问题。然后引入色彩相似性约束将灰度空间的匹配和彩色空间的匹配有机的结合起来,大大提高了匹配的精度。实验结果表明,该算法能够合理利用彩色立体像对中的色彩信息,从而生成准确度较高的视差图,是一种有效的立体匹配方法。

A novel stereo matching algorithm was presented. The constraint of color similarity was presented and used in stereo matching appropriately in this algorithm. Firstly, Sun of Squared Difference （SSD） method was improved by multiplying a weight template to minimize the effect of noise. Then, by adding constraint of color similarity, the grey space matching was combined with color space matching suitably and the improved SSD method could use the color information reasonably. The experimental results show that this algorithm can obtain more precise disparity map, which demonstrates the practical value of the algorithm.