天天看點

快速高分辨率圖像的立體比對方法Effective large scale stereo matching

《Effective large scale stereo matching》

In this paper we propose a novel approach to binocular stereo for fast matching of high-resolution images. Our approach builds a prior on the disparities by forming a triangulation on a set of support points which can be robustly

matched, reducing the matching ambiguities of the remaining points. This allows for efficient exploitation of the disparity search space, yielding accurate dense reconstruction without the need for global optimization.

       該文章發表在2010年的accv上面,看了一下公式和代碼都異常複雜,主要思想就是通過待比對圖像上的特征點作為比對的支撐點,在支撐點上做三角剖分,對視差在進行插值計算,但是效果一般,适用于實時性要求高的場合。

       整個工程都用OpenCV和openmp 完成,代碼異常複雜,但是整體上非常吻合工程師的代碼風格,可以學習一下,但是改進難度比較大。

效果:明顯看出效果和普通的全局優化算法還是有一定的差距。

快速高分辨率圖像的立體比對方法Effective large scale stereo matching

文章下載下傳連結:

<a target="_blank" href="http://www.cvlibs.net/software/libelas/">http://www.cvlibs.net/software/libelas/</a>

<a target="_blank" href="http://www.cvlibs.net/publications/Geiger2010ACCV.pdf">http://www.cvlibs.net/publications/Geiger2010ACCV.pdf</a>

代碼下載下傳連結:

<a target="_blank" href="http://www.cvlibs.net/download.php?file=libelas.zip">http://www.cvlibs.net/download.php?file=libelas.zip</a>

openmp版本:運作時間非常短大概0.1s左右完全滿足實時性要求。

<a target="_blank" href="http://www.cvlibs.net/download.php?file=libelas_omp.zip">http://www.cvlibs.net/download.php?file=libelas_omp.zip</a>

我調試好的工程:

<a target="_blank" href="http://download.csdn.net/detail/wangyaninglm/9321831">http://download.csdn.net/detail/wangyaninglm/9321831</a>