图像过渡区直接提取与分割快速算法  

Fast Algorithm on Image Transition Region Direct Extraction and Segmentation

作　　者：闫成新[1] 孙亮[2] 

机构地区：[1]中国石油大学(华东)机电工程学院,山东东营257061 [2]中国石油大学(华东)化学化工学院,山东东营257061

出　　处：《郑州大学学报（理学版）》2006年第4期74-77,共4页Journal of Zhengzhou University:Natural Science Edition

基　　金：中国石油大学博士基金项目;编号Y040313;东营市科技攻关项目

摘　　要：以往的过渡区直接提取与分割算法速度较慢,抗噪性差.提出将一种局部复杂度参数用于图像过渡区提取与分割算法,实验结果表明,算法运算速度快,抗噪性好,且不存在小样本问题,易于硬件实现.Traditional transition region extraction and image segmentation methods have draw backs such as low computational speed, and sensitivity to noise. A novel local complexity based transition region extraction and segmentation method （LC-TREM） is presented. Experimental results demonstrate that LC-TREM is much faster than traditional local entropy based method, and LC-TREM has good ability to deal with noise. LC-TREM has no small sample problem, and is easy to be achieved by hardware.

关 键 词：图像分割 局部复杂度 过渡区 阈值 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]