说明：  建立一种可以区分输入图形的形态学算法，算法的输入是这三种图形的一种（这三个图形是对湖、湾、路段的图像细化后的结果），输出是对这三种图形的识别。假设这三种图形都是1个像素宽且完全连通的，然而它们出现的方向可以是任意的。
(Can establish a distinction between the morphology of the importation of graphics algorithm input is a graph of these three (the three graphics on the Lake, Bay, after the section of the results of image thinning), the output is the three identification graphics. The assumption that these three are all graphics and a fully-connected pixels wide, but the direction they appear to be arbitrary.)