中文摘要：

针对传统局部立体匹配算法在深度不连续区域和低纹理区域匹配精度不高的问题，提出了一种基于SIFT描述子的自适应聚合权重立体匹配算法．算法首先采用梯度域的幅值和相位获取初始匹配代价；然后利用相似性区域判决准则获得各个中心点的自适应矩形聚合窗口，并利用各点SIFT描述子的L1范数进行自适应聚合权重计算．仿真实验结果表明，该算法能够有效地提高低纹理区域和深度不连续区域的立体匹配精度，获得较高精度的视差图．

英文摘要：

Aiming at the problem of low matching accuracy in both depth discontinuities and low textured regions of traditional local stereo matching algorithms,an adaptive support-weight stereo matching algorithm based on SIFT(scale-invariant feature transform)descriptors was proposed.First,the original matching cost was calculated with gradient amplitude and gradient phase.Then the adaptive support-window for each pixel was obtained with decision criterion based on the color similarity principle.Finally,with the L1 norm of SIFT descriptors of each pixel,the adaptive support-weight was calculated.Simulation experimental results show that the proposed algorithm can effectively improve the accuracy of stereo matching algorithms in both depth discontinuities and low textured regions,thus achieving higher matching accuracy.