基于模糊逻辑和区域合并的分水岭变换图像分割  

Watershed transform based on fuzzy logic and region merging algorithm

作　　者：林振荣[1] 黎嘉诚 刘九畅 俞鹏飞 李建民[1] LIN Zhenrong;LI Jiacheng;LIU Jiuchang;YU Pengfei;LI Jianmin(Department of Computer Science&Technology,Nanchang University,Nanchang 330031,China;School of Information Engineering,Nanchang Hangkong University,Nanchang 330063,China)

机构地区：[1]南昌大学计算机科学与技术系,江西南昌330031 [2]南昌航空大学信息工程学院,江西南昌330063

出　　处：《南昌大学学报（工科版）》2018年第1期88-92,共5页Journal of Nanchang University(Engineering & Technology)

基　　金：江西省科技支撑资助项目(20122BBE500049)

摘　　要：通过分析分水岭变换对图像细微变化敏感而产生过分割现象的缺陷,提出一种基于模糊逻辑的分水岭变换区域合并算法。首先使用形态学滤波方法对图像进行平滑降噪处理,并进行分水岭变换。再根据相邻区域的边缘有效性与灰度相似性,通过模糊推理及去模糊化方法获得区域可合并度。最后根据可合并度进行区域合并,获得分割结果图像。实验结果的数据分析表明:该方法在保证了重要轮廓信息不丢失的前提下,克服过分割现象,大量减少无效区域,提高了分割结果对图像中物体轮廓的概括性,较好地达到分割图像的目的。A watershed region merging algorithm based on fuzzy logic was proposed through the analysis of the over\|segmentation defect in watershed transform resulting from its sensitivity to subtle changes.Firstly,the morphological filtering was used to smooth and denoise the image,and the watershed transform was performed.Then,the region merging degree can be obtained by fuzzy reasoning and defuzzification according to the edge validity and grey\|level similarity of neighboring regions.Finally,the regions were merged according to the degree of merging,and the segmentation result image was obtained.Experimental results and data analysis showed that this method averted the excessive segmentation and reduced invalid areas in large numbers in the premise that ensuring the presence of significant contour information.It was a better way to improve the generality of object outline in segmentation results and achieve the aim of image segmentation.

关 键 词：图像分割 分水岭变换 模糊逻辑 区域合并 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]