于形态学梯度重建的分水岭分割

提出一种基于形态学梯度重建的分水岭图像分割方法.该方法在形态学梯度图像的基础上,利用形态学开闭重建运算对梯度图像进行重建,在保留重要区域轮廓的同时去除了细节和噪声.避免了标准分水岭存在的过分割现象及传统形态学开闭运算先平滑原始图像,后进行分水岭变换而造成的区域轮廓位置偏移.仿真实验证明,无论从消除过分割还是区域轮廓定位等性能方面,该方法均具有较好的分割效果.整个分割过程无需进行分割后的区域合并处理,降低了分割的复杂性;且分割过程只需选择合适的结构元素大小,增强了算法的灵活性.

Watershed Segmentation Based on Morphological Gradient Reconstruction

A method of watershed segmentation based on morphological gradient reconstructing was proposed.First,morphological gradient image was obtained.Second,opening and closing by reconstruction operators were employed to reconstruct gradient image,and then standard watershed transformation was used to implement image segmentation.Through gradient image reconstruction,important region contours are preserved while most small regular details and noise are removed.The over-segmentation of the standard watershed segmentation is avoided and the region contours bias caused by the traditonal method with opening and closing to smooth original image followed by the standard watershed segmentation is also eliminated.Simulations show that the method can efficiently eliminate over-segmentation,and hold the position of region contours without evident bias,also the method can produce good segmentation without post-segmentation processing,which in some degree reduces the complexity of the segmentation.The whole segmentation needs only one parameter,disk-shaped structure element,which is convenient to adjust to different applications.