【摘要】 在对各种文本定位与提取方法分析比较的基础上,针对标志牌文本的具体特点,提出了一种基于连通区域的标志牌文本提取方法———在灰度图像中定位标志牌区域,二值图像中提取标志牌文本,并用VC++编程实现对采集的自然场景下的标志牌文本图像进行了实验,结果证明,该方法能够准确定位并提取自然场景下的标志牌文本,获得文本字符。更多还原

【Abstract】 Based on the analysis and comparison of various methods locating and extracting sign text,we present an effective connecting area-based method to extract sign texts according to its special characteristics,which is locating sign areas in gray images and extracting sign texts in binary images. Then,we programmed in Visual C++ and did experiments on sign text images captured in nature scene.Experimental results illustrate that the proposed method can accurately locate the sign area and extract the texts in natural scene images, and then get text characters.更多还原