复杂背景下钢材编号提取与字符分割技术  

Extraction of Steel Number and Character Segmentation under Complex Background

作　　者：李俊[1] 陈建勋[1] 熊文龙[2] 万来毅[1] 

机构地区：[1]武汉科技大学计算机科学与技术学院,武汉430081 [2]武汉理工大学计算机科学与技术学院,武汉430063

出　　处：《武汉理工大学学报（交通科学与工程版）》2006年第5期915-918,共4页Journal of Wuhan University of Technology（Transportation Science & Engineering）

摘　　要：提出了一种复杂背景下多钢块图像编号的提取和分割方法.采用HSV颜色空间,并以饱和度为主区分前景和背景,再综合运用数学形态学运算及边界跟踪技术提取出钢块的有效区域.运用多投影直方图分割法对钢材区的编号进行行、列分割,并解决了干扰、字符粘连等问题,实现了编号的精确定位分割.实践表明该方法的分割效果是良好的.An effective method for steel number extraction and segmentation under complex background is presented in this paper. Based on the HSV color model, the method regards the saturation as principle to distinguish the foreground and background, then uses mathematic morphological operation and boundary detection technique to extract valid district of steel pieces. The multi-projection histograms are proposed to segment the code in the line and row, and it also solves interference and conglutination problems, achieves accurate location of the number. The practical results show that the segmentation method is good.

关 键 词：饱和度 分割 定位 提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]