摘要
为了降低图像噪声对分水岭算法产生的过分割的影响,在分水岭算法之前进行滤波预处理.而滤波算法在平滑噪声的同时会丢失图像的边缘结构信息,为了保持在滤波预处理过程中图像的边缘结构信息,使用相似性度量的滤波函数,并且相似性度量从基于像素点的方法扩展到基于图像块的方法.
In order to reduce the impact of over-segmentation of the watershed algorithm for image noise.we do the filter preprocessing first. The noise filtering algorithm, while smoothing the edge structure information of the image will be lost. In order to maintain the edge structure information in the image pre-processing filtering using the filter function similarity measure, and similarity measure extended from pixel-based approach to the method based on image block.
出处
《喀什师范学院学报》
2014年第6期45-49,共5页
Journal of Kashgar Teachers College
关键词
分水岭
预处理
边缘保持
Watershed
Preprocessing
Edge retention