Modified Watershed Algorithm for Image Segmentation
In order to solve the problem of over-segmentation existing in traditional watershed algorithm, an improved segmentation algorithm based on watershed is proposed.Firstly, the reconstruction filter based on opening and closing is adopted to remove the noise of images.Then,multi-scale gradient operators are used to calculate morphological gradient of the filtered images, which can overcome the shortcomings of other conventional gradient operators and produce an accurate gradient image suitable for watershed segmentation.Finally, get the reconstruction morphological gradient images of the different dimensions structural elements and calculate its deep local minima binary maker image in the specified threshold by using improved marker extraction algorithm.The union of a series of binary marker images is regarded as the final marker image.The markers are used to modify morphological gradient image and get a final split line.Compared with the traditional watershed method, the final result shows that this method can not only effectively remove the noise and avoid the over-segmentation of watershed, but also enhance the robustness of segmentation.
Image segmentation Watershed algorithm Morphological gradient Marker extraction
SUN Fengli WANG Xichang
Department of Management Science and Engineering,Shandong Normal University,P.R.China 250000 Shandong Province Department of Education,P.R.China,250000
国际会议
山东 日照
英文
15-20
2015-08-17（万方平台首次上网日期，不代表论文的发表时间）