摘要
传统的分割方法往往是根据图像单一的属性标准对图像进行分割,很难满足图像的多方面分割要求。由于许多外界干扰因素的存在,使得基于经典模糊集方法进行的分割,结果也常常不令人满意。针对这些问题,介绍了一种基于超模糊集合理论的多属性图像阈值分割方法(F2ES),在超模糊集的基础上,结合模糊熵和模糊相似性两种截然不同的属性刻画待分割图像,构造综合评价函数,得到最佳阈值。针对多幅不同类型图片进行分割仿真实验,得到较好的结果,证明该算法是切实可行的。
Classical measures partition one image according to a single property. So, it is difficult to satisfy requests of the image thresholding. Furthermore, due to disturbing factors, the result of image thresholding based on fuzzy sets is not always satisfactory. A new thresholding using multi-properties based on ultra-fu~ sets was proposed （F2ES）, which processed optimal threshold as comprehensive assessment function constructed by fuzzy entropy and fuzzy similarity based on ultra-fuzzy sets. Experimental results and simulations conducted on various images are provided to testify the validity of the proposed algorithm.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2007年第19期4434-4439,4444,共7页
Journal of System Simulation
基金
中国国家自然科学基金(60274099)
关键词
图像阈值
模糊集
超模糊集
模糊熵
模糊相似性
综合评价函数
image threshold
fuzzy sets
ultra-fuzzy sets
fuzzy entropy
fuzzy similarity
comprehensive assessment function
