基于蚁群优化算法的复杂背景图像文字检测方法  

Text detection from images with complex background by ant colony optimization algorithm

作　　者：李敏花[1] 柏猛[1] 

机构地区：[1]山东科技大学电气信息系,济南250031

出　　处：《计算机应用》2011年第7期1844-1846,共3页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61071087);山东科技大学春蕾计划资助项目(2010AZZ0502010AZZ049)

摘　　要：针对复杂背景图像中的文字检测问题,提出一种基于蚁群优化算法的复杂背景图像文字检测方法。该方法首先采用蚁群优化算法提取图像边缘;然后在边缘图像上提取特征,采取由粗到精多级检测、验证的策略进行文字检测。与基于Soble算子、Canny算子等方法的对比实验结果表明,所提出的基于蚁群优化算法的文字检测方法可有效地实现复杂背景图像中的文字检测。To detect text from images with different backgrounds,a text detection method with ant colony optimization algorithm was proposed.Before text detection,an ant colony optimization algorithm was adopted to detect image edges,and then features were extracted from the edge image.Afterwards,a coarse-to-fine strategy was applied to detect the text lines from image.Finally,the experimental results show that the proposed method achieves more precise detection than Soble-based,Canny-based and othor two detection methods.

关 键 词：文字检测 边缘检测 蚁群算法 复杂背景 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]