改进的非参数Census变换立体匹配算法

针对多数立体匹配算法的相似性测度都建立在像素灰度特性基础上, 无法彻底消除匹配差异性, 易出现歧异性的问题, 提出一种改进的非参数Census变换匹配算法. 该算法通过在传统非参数Census匹配过程中增加局部纹理反差值测度, 引入图像纹理度量的方向性, 使中心像素灰度值不再是唯一决定因素, 改进了匹配模版, 从而有效解决了传统匹配算法的歧异性问题. 实验结果表明, 改进算法是一种有效、 合理的立体匹配方法, 提高了稠密匹配精度.

Improved Stereo Matching Algorithm Based onNonparametric Census Transform

Aiming at the problem that the similarity measure of most stereo matching algorithm was mostly based on gray characteristics of pixels, which could not completely eliminate the differences of matching, and was prone toambiguity, we proposed an improved matching algorithm based on nonparametric Census transform. It increased the measure of local texture contrast value in thetraditional nonparametric Census matching process, and the directionality of theimage texture was introduced, so that gray value of the center pixelwas no longer the sole determinant, and matching template of the transform modewas improved, thus the ambiguity problem of traditional matching algorithm waseffectively solved. The experimental results show that the improved algorithm is an effective and reasonable method of stereo matching, which improves the accuracy of dense matching.