复杂场景下基于Radon校正与MSER的文本定位算法研究  

Research of text localization algorithm based on Radon tilt correction and MSER in complex scenes

作　　者：刘杨[1] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221116

出　　处：《微型机与应用》2016年第21期42-44,48,共4页Microcomputer & Its Applications

摘　　要：针对复杂场景下,由于文字背景复杂性而带来的文字定位不准确的问题,提出了一种Radon倾斜校正与MSER(最大稳定极值区域)结合的算法,基于该算法识别出文字在所处复杂背景中的位置。首先,对图像进行预处理,采用Canny算子对图像进行边缘检测,并利用Radon变换对倾斜图像进行校正。然后通过二进小波变换对图像进行增强处理,去除噪声。最后提取MSER,经过粗处理、细处理后,检测出文本所处位置。实验结果表明,基于Radon倾斜校正与MSER结合的方法,能够有效提高文本定位的准确率。For complex scenes,due to the complexity of the background of text,text positioning is not that accurate. So the Radon tilt correction algorithm combined with the MSER（ Maximally Stable Extremal Region） was proposed. Based on the algorithm,text location in complex background can be identified. Firstly,the image pretreatment was carried out,the Canny operator was used to detect the image edge,and the Radon transformation was performed to do the image tilt correction. And then the image was enhanced and noise was removed through dyadic wavelet transformation. Finally MSER should be extracted,and after the processing of coarse and fine processing,the character position will be detected. The experimental results show that the method based on Radon tilt correction and MSER can effectively improve the accuracy of text positioning.

关 键 词：Radon倾斜校正 CANNY算子 二进小波变换 最大极值区域 文本定位 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]