摘要
地形图中包含大量的汉字注记,为了实现地图要素自动识别,首先要将文字正确地提取出来。该文通过对汉字特征的分析,提出了一种新的汉字提取算法。首先根据地图图像的特点进行骨架运算,利用形态算子对汉字笔画进行分解,最后实现对汉字注记的笔画统计提取。该算法较好地解决了字线粘连、旋转等情况下的汉字提取问题,具有良好的稳定性。
There are plenty of Chinese characters in topographic maps.They must been extracted correctly before the automatic recognition of map elements.A new effective algorithm is presented based on the feature analysis of Chinese characters.First,the skeleton algorithm is implemented according to the feature of map images.And then morphological operators are used to decompose characters.At last the characters are extracted with the method of analyzing the statistical strokes.The algorithm could extract characters of adhesion or rotation effectively.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第3期11-13,27,共4页
Computer Engineering and Applications
基金
部委"十五"项目基金资助
