不规则三角网高程的遮挡检测算法

Occlusion detection algorithm based on TIN elevation

针对多视影像匹配中由于遮挡导致的问题,该文提出一种基于不规则三角网(TIN)高程的遮挡检测算法。该方法根据已知的高程格网平面,构建出物方不规则三角网;根据三角网内插得出某一个地面点的高程值。对于这一地面点,如果在其搜索路径上的搜索点高程均低于搜索点所在位置的投影光束高度,则该点必定在影像上可见。实验表明,该文提出的算法能高效、准确地得到不可见的点,特别对尺寸较大的影像能快速计算。

In order to solve the problem caused by occlusion in multi-view image matching,an occlusion detection algorithm based on triangulated irregular network（TIN）elevation was proposed.According to the known elevation grid,a triangular irregular network of objects space was constructed,based on which elevation of a certain ground point was obtained by interpolation algorithm.For this ground point,the point must be visible on the image if the search point elevation on its search path is lower than the projection beam height located at the search point.Experimental results showed that the algorithm proposed in this paper can get invisible points efficiently and accurately,especially for larger scale images.