针对文本图像拍摄质量低下,而导致OCR系统识别率不稳定的问题,本文提出了一种基于文字笔画结构的文本图像校正算法,主要是在图像的预处理中对待处理像素点进行基于文字笔画结构的特征分析,实现目标和背景像素的校正,再结合局部二值化算法进行处理,分析其噪音分布特点采用邻域去噪进一步优化处理结果。实验表明,本文算法能够很好的适用于处理质量低下的文本图像,处理效果从视觉图像和识别率上都能满足应用需求。
Given the facts that OCR system recognition is unstable because of the low quality text image, this paper put forward a text image correction algorithms base on the structure of text strokes. The algorithm is mainly to analysis the characteristics of the processing pixel based on the text strokes structure, in order to correct the goals pixel and background pixel. Then, combined with local binary processing algorithm for processing, and using neighborhood de-noising method to further optimized image de-noising , The experimental results show that the method can be applied to deal with poor quality text images, and the visual effects and image recognition rate can meet the application.