摘要：针对传统的DP算法存在条纹瑕疵，提出了一种改进的垂直性约束动态规划立体匹配算法，通过利用扫描线间的相关信息来约束动态规划过程，使得处理每一条扫描线时都能充分利用以前扫描线的匹配信息。实验结果表明，该算法可以明显地消除传统DP算法的视差图的条纹瑕疵，匹配精度高, 而且执行时间短。-ABSTRACT Although network speed has been increased over the years,bandwidth limitation continues to pos

利用不变矩公式将双目灰度图像转换成双目不变矩图像，然后提出了具有自适应参数的简单树动态规划算法的双目立体匹配方法进行匹配得到最终视差图-The formula using invariant moments grayscale images into binocular binocular image invariant moments, then made ​ ​ a simple tree with adaptive parameters of the dynamic p

利用动态规划和互信息计算初始视差图，在此基础上对初始视差图进行左右一致性检验等进行求精(The initial disparity is computed according to dynamic programming and Mutual Information. Furthermore, the consistency examination is utilized for disparity refinement)