气象传真图图题的识别  

Recognization of the title in meteorologic fax image

作　　者：刘繁明[1] 孙铭雪 邢珅 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《应用科技》2010年第12期52-55,共4页Applied Science and Technology

摘　　要：针对目前气象传真图种类识别人工干预,执行效率低的问题,提出了一种通过识别气象传真图图题来判断传真图种类的方法.首先获取图题在整幅气象传真图中的位置,然后利用投影、轮廓切分分离粘连字符方法完成对气象传真图图题字符的分离,最后通过提取字符特征量,采用模板匹配算法完成字符识别,从而确定出气象传真图的种类.该方法实现传真图类别的自动识别,避免人工干预,自动化程度高,应用于气象传真图数字处理系统,取得较好的识别效果.In allusion to the problem that the recognization of meteorologic fax images＇ category still depends on manual intervention and the efficiency is lower,the authors brought forward a new way to judge the fax images＇ category by recognizing the title of meteorologic fax image.The method first got the position of the image title according to the whole image,then used projection and syncopated conglutinated characters based on contour to accomplish the separation of the characters in the image title,and finally used templet match to recognize the characters by distilling characteristics of the characters.This method realizes automatic recognizing of the category of the fax images and avoids manual intervention,reaching a high automatic degree.This method is used in meteorologic fax image numeric disposal system and achieves preferable effect.

关 键 词：气象传真图 粘连字符 字符识别 模板匹配 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]