基于模糊C均值的图像目标检测技术研究  

Detecting of image objects based on FCM algorithm

作　　者：王占江 

机构地区：[1]中国人民解军92515部队

出　　处：《信息技术》2014年第3期5-8,共4页Information Technology

基　　金：国家自然科学基金(10771043);水下机器人国防技术重点实验室基金(002010260730)

摘　　要：对经典FCM算法的优缺点及其改进算法进行了综述,在此基础上,构造了基于概率密度和基于模拟退火与粒子群相结合的核函数聚类两种新算法。仿真结果表明:两种算法都具有较好的图像分割效果,尤其方法 2运算速度快,且具有一定的全局搜索能力。该综述对于丰富图像分割研究具有一定的参考价值和指导意义。For the FCM algorithm in the application of image segmentation, these advantages and disadvantages of the FCM algorithm and the typical improvement algorithm are reviewed. On the basis of the review is constructed based on the probability density and integration based on SAPSO and KFCM two new algorithms. The results show both algorithms are good segmentation results, especially method 2 fast speed, and with some global search capability. Summary for the segmentation of rich has some reference value and significance.

关 键 词：模糊C均值 K 概率密度 图像分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]