一种基于局部特征的拓片图像边缘检测算法  

A New Edge Detecting Algorithm Based on Local Characteristic in Rubbings' Processing

作　　者：路晓菊[1] 张太镒[1] 张亚歌[1] 董俊伶 

机构地区：[1]西安交通大学电子与信息工程学院,陕西西安710049 [2]新疆昌吉电业局科技信息部,新疆昌吉831100

出　　处：《微电子学与计算机》2010年第6期97-100,共4页Microelectronics & Computer

摘　　要：针对传统边缘检测算法在处理拓片图像时,容易将模糊的细节部分丢失或者误检的问题,提出了一种基于局部特征的拓片图像边缘检测算法.该算法预处理部分采用了unsharp算子对细节部分进行保护和增强,利用局部标准差均值比(Local Standard Deviation Mean Ratio,LSDMR)对图像进行边缘描述,对LMSDR进行处理以得到图像边缘,实验结果表明,该算法对拓片图像模糊细节的边缘检测明显优于传统算法.In this paper,a new edge detecting algorithm based on local characteristic in rubbings＇ processing is propsed.Compared with traditional edge detecting method,the new one have better performance in detecting fine structures.In the new algorithm,unsharp operator is used in pretreatment to protect fine part and increase contrast,then we describes edge in picture using Local Standard Deviation Mean Ratio（LSDMR）of pixel,after processing of LMSDR we get rubbing＇s edge.Experiments show the effectiveness of the new algorithm obviously.

关 键 词：拓片 局部特征 局部标准差均值比 边缘检测 图像处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]