基于梯度值的自适应窗口立体图像匹配算法

针对基于区域的立体图像匹配算法支持窗口难以选择,容易出现窗口过大或过小的问题,提出一种新的自适应窗口立体图像匹配算法。该算法利用Sobel梯度算子计算像素梯度值,并根据其梯度值动态地获取具有自适应的支持窗口,然后分别选择相似性测度函数SAD或NCC搜索最佳匹配点,获得视差图。此外,算法在窗口选择过程中进行优化,减少了计算量。实验结果表明,改进后的算法提高了匹配正确率且计算时间缩短了近5%。

Stereo Matching Algorithm of Adaptive Window Based on Gradient

Aiming at the problem that the support windows are selected difficultly in stereo matching algorithm based on local areas and the support windows are bigger or smaller, a new stereo matching algorithm of adaptive windows is proposed. First, the algorithm acquires size of support windows adaptively by computing gradient of pixel with Sobel operator, then searches the most ap- propriate pixel to get disparity image by function of SAD or NCC. In addition, the method is improved by decreasing the computation in selecting size of windows. The results show that this algorithm has higher precision and its calculating time is shortened by nearly 5%.