首页 | 本学科首页   官方微博 | 高级检索  
     

基于高斯拟合的自适应多阈值分割算法
引用本文:李磊,平西建,童莉.基于高斯拟合的自适应多阈值分割算法[J].计算机应用研究,2006,23(8):162-163,170.
作者姓名:李磊  平西建  童莉
作者单位:信息工程大学,信息工程学院,河南,郑州,450002;信息工程大学,信息工程学院,河南,郑州,450002;信息工程大学,信息工程学院,河南,郑州,450002
摘    要:提出一种新的基于邻域平均直方图的快速高斯拟合函数参数估计的算法,对形态滤波后的邻域平均直方图进行峰谷检出,对各峰区数据单独进行截断数据统计分析确定拟合参数,较好地避免了多峰交叠所带来的干扰,从而实现了复杂图像的自适应多阈值分割。实验表明,该算法能有效提高拟合精度和分割的鲁棒性。

关 键 词:图像分割  邻域平均  高斯拟合  参数估计
文章编号:1001-3695(2006)08-0162-02
收稿时间:2005-07-13
修稿时间:2005-07-132005-09-23

Adaptive Multi-method Segmentation Algorithm Based on Gaussian Fitting
LI Lei,PING Xi-jian,TONG Li.Adaptive Multi-method Segmentation Algorithm Based on Gaussian Fitting[J].Application Research of Computers,2006,23(8):162-163,170.
Authors:LI Lei  PING Xi-jian  TONG Li
Affiliation:Institute of Information Engineering, University of Information Engineering, Zhengzhou Henan 450002, China
Abstract:To alleviate these undesirable disturbance between peaks,and fit multi-peak feature of histogram more accurately,a new and fast estimation algorithm of Gaussian fitting functions'parameters based on local average histogram is introduced in this paper.After morphology filtering,peak and valley are detected in the histogram.Statistical theory for truncated data ana-(lysis) are applied to estimate the parameters of fitting function of each peak area,and get optimum thresholds.Experimental results show that this algorithm can be automatically and effectively applied for the multi-threshold segmentation.
Keywords:Image Segmentation  Local Average  Gaussian Fitting  Parameter Estimation
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号