The Research Of Correcting Method For Warped Chinese And English Mixed Document Images

In this digital age, digital image processing technology has been used in many industries. With the development of OCR (Optical Character Recognition) technology, a lot of conveniences have been brought to people’s life. Before the OCR processing, because of the problem such as illumination unbalance, book tilted, warped text line due to the thickness of book, can cause the character recognition rate of OCR reduced. To resolve this problem, distortion correction should be used to the image before the OCR processing. The existing correcting methods are mostly based on single language, they all have limitations to the languages mixed document images.This paper will address the distortions in the Chinese and English mixed document image. Through the research of characteristic in this kind of image, summarizes the results of existing distortion correction for Chinese or English text images, comparative the advantages and disadvantages of various methods, a fast distortion correcting method for warped Chinese and English mixed document images is proposed in this paper.The development of correcting methods for warped document images and the research contents of this paper are introduced at the beginning of the thesis. Then it shows the ideas of this algorithm. The second part introduces some techniques for image preprocessing. After that, the article analyzes the characteristics of warped Chinese and English mixed document images. Algorithm in this paper is proposed based on the research, and the practicability is analyzed. Then it introduces the implementation of this correcting method. The introduction focus on the text line positioning and character segmentation. Distortion correction is realized by each module which introduced before. Next part is the evaluation of the algorithm in this paper. The experiment results are analyzed by the references of OCR recognition rate and correction time. The correcting method is summarized at the end of this thesis.Experiments show that this correction could rectify the warped Chinese and English document image quickly and effectively. The OCR rate of the corrected images could be significantly improved. So this is a distortion correcting method with good application prospect.