基于数学形态学的遥感影像水域提取方法

A new algorithm for water area extracting from the remote sensing image based on morphologic method

利用数学形态学的方法对遥感影像上的水体进行了提取实验。经实验结果得出,利用数学形态学的基本运算组合成的各种算子:高、低帽变换、滤波与阈值截取相结合的增强方法,通过选择合适的结构元素,对图像进行预处理后进行分割,可以完成遥感影像上灰度变化较小的水体提取,其结果优于直接对原图像进行形态梯度法、均值分割的结果,对分割后的结果进行矢量跟踪后,得出的结果可以直接应用到GIS中去。

In this paper,water area extracting experiments are made on RS image by Morphologic methods. Based on the combination of all kinds of basic Morphologic operators which leads to different top and bottom Hat transformations, filtering and thresholding enhancement are used to pre-process the image, by choosing appropriate structure elements,the image is segmented to complete the water area extracting with small gray variation.The results shows that,the method presented in this paper works better than the direct segmentation methods such as morphologic gradient method and mean method. The segmented image can be directly used in GIS after vector tracking.