There is the abuse from the traditional watershed segmentation algorithm,which is caused by extracting the object points less accurately.This paper used interactive segmentation,combined with the method of pre-filtering to further modify the gradient image,and used stacking dam to enhance the corrected images.The paper did some tests such as extracting points interactively,pre-filtering and correcting segmentation through making use of the C++ programming language and the queue data structure to store the pixels.The test results showed that the algorithm of this paper greatly reduced or eliminated the oversegmentation,and improved the accuracy of segmentation.%针对传统水域分割算法提取目标点不准确而造成过分割的弊端,采用交互式分割,同时结合图像预滤波的方法进一步修正梯度图像,并对处理后的图像使用叠坝算法进行增强.本文使用C++语言编程,使用队列数据结构进行像素存储,进行了交互取点、预滤波及修正分割等试验,试验结果表明本文算法极大程度上减少或消除了图像的过分割现象,提高了分割的准确度.
展开▼