【摘要】 流域变换是图像分割的有力工具.流域分割方法的性能主要依赖于图像梯度.文中利用传统单尺度梯度算子流域分割算法和多尺度形态梯度流域变换算法,分别作用于计算机视觉细胞图像和水果图像,结果显示传统分割算法和多尺度算子流域分割算法在不同的图像中显示出各自的优越性,第一种算法更适合于交叠区域不是很多,但质地不必十分均匀的图像;第二种则更适合于交叠区域较多,质地较均匀的图像.更多还原

【Abstract】 Watershed transformation is a powerful morphological tool for image segmentation. The performance of segmentation methods based on watershed depends largely on the gradient of the image. The conventional transformation and the new one of watershed based on multi-scale gradient operator are discussed in this paper. They are used in cell and fruit images respectively. The results demonstrate the two methods have their own advantages respectively.The first one works better with an image of less overlapped objects in the irregular back-ground. On the contrary, the second one is well applied to images of more overlapped objects in regular background.更多还原