显微图象中目标对象的边缘提取研究  

Study on Edge Extraction of Objects in Microscopic Images

作　　者：王跃宗[1] 刘冲[1] 王立鼎[1] 徐征[1] 

机构地区：[1]大连理工大学机械工程学院,辽宁大连116024

出　　处：《小型微型计算机系统》2003年第12期2177-2180,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金 (50 2 750 2 3)资助 ;重庆市快速成型与模具寿命重点实验室项目 (CQMK 1 0 )资助

摘　　要：研究了显微图象中目标对象边缘的提取 ,提出了一种新方法 .该方法分三步进行 ,首先识别目标对象 ,并与背景区分开 ,然后滤除误识别象素和脏点 ,最后提取目标对象的边缘 .使用最优颜色通道识别显微图象中的目标对象 ,动态的调节识别阈值 ,减少了误识别象素的数量 .对残存的误识别象素和脏点采用面积滤波的方式去除 ,进一步提高边缘提取的效果 .试验结果表明 ,采用本文的方法提取目标对象的边缘优于已有的熵算符法提取的结果 .A new method is applied to extract edges of object in microscopic images.Three steps are carried out. Firstly, the object is recongnized from background .Secondly, the erroneous and dirty pixels are filted. Thirdly, the edges of object are extracted . The erroneous pixels are reduced by choosing the best color channel to recongnize object and dynamicly adjusting the recongnized threshold. The level of edge extraction will be improved ulteriorly by area filtering erasing the remained erroneous and dirty pixels. Experiment results indicate that the method of edge extraction in this paper outperform entropy operator.

关 键 词：显微图象 边缘提取 阈值 象素 图象处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]