复杂背景下文字增强算法研究与应用-通信与信息系统专业论文.docx

第 第 PAGE IV 页 第 第 III 页 理取得了很好的效果。采用本文提出的的二值法算法，准确度和清晰度都 非常高，能大大提高文字的识别率，其效果明显优于其它方法。 最后，基于两种新的二值化算法的相关实验结果，得出优缺点和适用 范围，为今后的图像增强算法的研究提供了参考的依据。与此同时，本文 在文字增强模块后级联了文字识别测试系统，可以快速准确地把图像中的 文字识别出来。 关键词：文字增强算法，复杂背景，退化图像，二值化，文字识别 Novel Character Enhancement and recognition System in complex background ABSTRACT Multimedia data contains huge character information. Text enhancement and recognition is a vital and fundamental problem of processing images. In pace with the abundantly representative element and complex background, recognizing the characters in the image becomes more difficult work. In actual application, the character information is very important. They can rapidly and exactly deliver what the picture whats to say. So, extracting this information means too much. Firstly, this paper discuss the main backyards of character enhancement and recognition knowledge and analyze the popular technology in different level and feilds. Based on this, the paper introduce the primary object which is being studied. That is the binarization technology. Secondly, the paper analyze many popular binarization algorithms. We not only know the theory, but also program to achieve them in many experiments. By comparing them with the advantages and disadvantages, we can find that they are not suitable to the complex background information extracting work. Thirdly, the paper has proposed two innovations based on the existing technology. The first algorithm mainly puts focus on the IC chip whose surface is sensitive to the strength and angle of the light. Those degraded image can be well processed by the first algorithm. The second method is not only adapted to different color, font, language, but also getting rid of much non-text information and automatically changing the text to black and other white. Compared with other binary method, this one can highly raise the text recognition rate. At last, based on the two new binarization system, the paper has analyzed the advantages ,disa