自适应窗口快速立体匹配

Fast Stereo Matching Using Adaptive Window

灰度立体匹配的关键问题是选择一个合适的匹配窗口,该文对Kanade等提出的通过测算局部灰度和视差变化来选取适当窗口的方法进行了改进,提出一种仅利用灰度信息即可确定窗口的简单方法.该方法在能够求得最佳窗口的前提下,可以方便地利用整数图像来提高效率,计算量与窗口大小无关.在国际标准图像上测试的实验结果表明该方法在局部匹配方法中是最优的,虽与有些全局方法相比有一定差距,但其计算复杂度远远低于它们.

It is crucial to choose an appropriate window size for area-based stereo matching. A method that only use gray information to decide the window size is presented in this paper, which improves the approach using local gray and disparity variation to select window size. On the premise of a sure find of an optimal window size, this method can easily make use of integral image to speed up matching computation, so the calculation is independent of the window size. The proposed method is tested with the international standard image data and compared with some other methods. The experimental results show that its performance is the best among local matching methods, although worse than some excellent global-optimal approaches, its complexity is much lower than them.