Matlab的标记分水岭分割算法
1 综述 Separating touching objects in an image is one of the more difficult image processing operations. The watershed transform is often applied to this problem. The watershed transform finds "catchment basins"(集水盆) and "watershed ridge lines"(山脊线) in an image by treating it as a surface where light pixels are high and dark pixels are low. 如果图像中的目标物体是连接在一起的,则分割起来会更困难,分水岭分割算法经常用于处理这类问题,通常会取得比较好的效果。分水岭分割算法把图像看成一幅“地形图”,其中亮度比较强的区域像素值较大,而比 较暗的区域像素值较小,通过寻找“汇水盆地”和“分水岭界限”,对图像进行分割。 Segmentation using the watershed transform works better if you can identify, or "mark," foreground objects and background