如果图像中的目标物体是连在一起的，则分割起来会更困难，分水岭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副“地形图”，其中亮度比较强的地区像素值较大，而比较暗的地区像素比较小，通过寻找“汇水盆地”和“分水岭界限”，对图像进行分割。-If the object in the image together, then split up and it will be more difficult, the watershed algorithm often used to deal with such issues, usually obtain better results. Watershed segmentation algorithm image as a topographic map, strong brightness region pixel value, while the dark areas are relatively small pixels, image segmentation by finding the catchment basin and watershed boundaries .