Research And Implementation Of Printed Chinese Character Recognition System For Smart Phones

With the progress and development of society, various information is increasing, the most common information is the printed Chinese character document. With the rapid development of computer technology, people are no longer satisfied with the use of manual operation to obtain the text information on the document. Optical character recognition technology can convert Chinese characters of images into computer texts. Optical character recognition technology can greatly save the time to obtain the document content and reduce unnecessary workload.Huge quantity is one of the important characteristics of Chinese characters. There are 3755 Chinese characters in first level which were frequently used .There were 3008 Chinese characters in second level. The number of total Chinese characters is between 80000-90000.At the same time, due to the unique development history of Chinese characters, the structure of Chinese characters is complicated, and there are many structurally similar words. Because of the large number of Chinese characters and complex structure, compared with English letters which are structurally simple, the recognition of Chinese characters is more difficult.Now the smart mobile phone is very popular, and smart mobile phone hardware is close to the level of personal computer, intelligent mobile phone can expand the application scenarios, so the application of Chinese characters recognition technology in smart mobile phone has practical significance. The researches on image preprocessing, image segmentation and recognition Chinese characters had been made. At the same time, many experiments for various identification schemes were did and the experimental results of different schemes were analyzed. Combining the width and height geometric features and the vertical projection method of the Chinese characters, this paper proposes a single character segmentation algorithm. Finally implement a practical and convenient printed Chinese characters recognition system on smart mobile phone. The research work in this paper includes:(1) Image preprocessing. The image processing methods were studied for gray scale,binarization and noise removal. Through a large number of experiments, the image binarization and denoising algorithms had been chosen for smartphone applications. This step can convert the original image to basic noise-free binary image.(2) Segmentation of the characters. Various algorithms for image line extraction, tilt correction and single character segmentation are studied. A single Chinese character segmentation algorithm is proposed based on Chinese characters’ feature. The experimental results show that the algorithm can cut the whole Chinese character as a whole.(3) Chinese characters recognition. Through the study of normalization, skeletonization and feature extraction of character images, it is shown that the feature extracted by grid is suitable for Chinese character matching. Finally, the advantages and disadvantages of this method were analyzed.(4) Design and implement a mobile application based on Android. According to the functional requirements of common OCR products, designed and implemented a printed Chinese character recognition system for Android system.