一种闭合曲线去除平面工件二值图像噪声算法  

An Algorithm of Eliminating Noise for 2D Workpiece of Binary Images By Closed Curve

作　　者：王中王[1] 薛建彬[1] 张学友[1] 

机构地区：[1]南京航空航天大学机电学院,江苏南京210016

出　　处：《工业控制计算机》2010年第11期55-56,共2页Industrial Control Computer

摘　　要：在现代工业中,图像处理技术经常被用于工件的在线检测。获取的图像经过二值化后,工件信息和噪声同时存在,其中工件信息常被噪声影响。通过分析二值图像中工件边缘与噪声的特征区别,提出一种用闭合曲线去除二值图像噪声的新算法,该算法通过移动一组闭合曲线,根据每个闭合曲线与检测目标轮廓边缘的交点数目,判断该组闭合曲线所包含的各个像素点是否为噪声,将被判为噪声的像素点置零,从而提取出精确的工件边缘信息。该算法适用于具有清晰轮廓曲线的零件检测。The image information of the workpiece is often expressed in binary images,where both the noise and the outline of workpiece exist.The noise will affect the detection accuracy of the workpiece outline.It＇s important to eliminate the noise of binary image.An algorithm was proposed to eliminate the noise using a closed curve method,after analyzing the different characteristics between the workpiece outline and the noise.A set of closed curves move continuously.Each curve will intersect with the workpiece outline at certain points.According to the number of intersection points,the pixels inside the closed curve are judged whether the points are noise or not.The noise pixels are eliminated by setting as zero.So a more accurate workpiece outline is obtained.

关 键 词：在线检测 二值图像 噪声消除 闭合曲线 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]