【摘要】 针对立体匹配当中的匹配精度问题,提出一种基于迭代指导滤波的立体匹配方法。该方法首先利用自适应窗口方法根据图像的灰度变化自适应地选择支撑窗口,再根据指导滤波累积原始匹配成本,并采用"胜者全取"方法计算视差,然后利用计算视差对匹配成本进行修正,进行再次累积并计算视差,最后反复迭代这一过程直到达到最大的迭代次数为止。实验结果表明,该方法可以获得较高精度的视差图,而且对立体像对中的不连续区域具有较强的鲁棒性。更多还原

【Abstract】 This paper presents a stereo matching method based on iterative guided filter aimed at improving the accuracy of stereo matching. In this method,an adaptive window method is used to make an adaptive choice of a support window according to grayscale variation in images. Secondly,raw matching cost is aggregated by a guided filter and disparities are computed by the "winner-take-all " method. Thirdly,matching cost is modified on the basis of the computed disparities and is again aggregated to calculate disparities. Finally,this process is repeated until the maximum number of iterations is reached. Experimental results show that the proposed method can achieve an accurate disparity map,and has a strong robustness to the discontinuous regions in images.更多还原