使用Bhattacharyya系数的道路自动提取

Automatic road extraction with Bhattacharyya coefficient

介绍了基于Bhattacharyya系数的小比例尺正射影像道路提取方法。首先假设道路点的灰度特征符合标准正态分布,建立理想道路模型,在影像中通过计算Bhattacharyya系数评价该位置为道路点的可能性,系数较大者为道路点。实验证明该方法具有一定的可行性。

In this paper,a method on road extraction from small scale orthoimage based on Bhattacharyya coefficient is introduced.First,the road is assumed as the set of the center point whose ideal gray distribution feature can be supposed as normal distribution（also known as Gauss distribution）,and then Bhattacharyya coefficient is calculated for each pixel of the candidate image.And the ones with greater coefficient values can be seen as road center points.Experiments prove the validity of this algorithm.