Method of Edge Extraction Based on Wavelet Transform for the Image of Mechanical Engineering Drawings  

基于小波变换的机械工程图扫描输入图像的边缘提取(英文)

作　　者：吴林瑞[1] 陈炯柏[1] 袁曾凤[1] 

机构地区：[1]北京理工大学机电工程学院,北京100081

出　　处：《Journal of Beijing Institute of Technology》2000年第3期335-340,共6页北京理工大学学报（英文版）

摘　　要：A method, by which the broken edge of mechanical engineering drawings being binarised can be eliminated and the whole edge of mechanical engineering drawing can be got, is given. To all points of a connected area on the image of the modular maximum value of wavelet transform at scale 2 2, the averaging grey value method is used to their grey values, then the edge of the dim place is continuous after the maximum variance threshold method is used. All these methods are fast, they can be used for all linear graphics having nothing to do with the grey value, its application scope is wide.给出了消除机械工程图扫描图像二值化产生的线的边界断裂、得到图像中完整边界的一种方法 .通过对第二层小波变换模极大值图像上的连通区域上各点的灰度 ,使用灰度平均方法进行平均 ,然后对第二层小波变换模极大值图像运用最大方差阈值法进行二值化 ,使图像中模糊之处的边界完整保留下来 ,保持了边界的连续性 .该方法速度快 ,适用于与灰度无关的线型图像 。

关 键 词：wavelet transform CONNECT FILTERING BINARIZATION 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]