Design And Implementation Of The Electronic Dictionary Based On Mobile Phone Camera

In recent years, with the rapid promotion of globalization, international communication is increasing faster and faster. How can rapidly, accurately and easily getting the Chinese meaning of an English word needs to meet urgently. On the other hand, with the popularity of smart phone which has a high resolution camera, realizing a real-time dictionary based on camera will become a very attractive application, which runs on the very popular android operating system.This paper designs and implements a real-time dictionary application based on android platform. Using the camera of mobile phone to get pictures of English text and part of handwriting words, through the method of pattern recognize and offline/online query, the application can give users the Chinese explanation of the English words. Studies of this paper focus on the existing optical character recognition method. System mainly includes the following three parts:Using the camera to take words; Optical character recognition; Local/Network translation. Specific research work includes:In order to get the translation, combining with the existing image processing algorithm[1], we divide image processing into preprocessing and standardizing. In preprocessing module, use the adaptive thresholding algorithm, which breaks the disadvantages of fixed threshold. In combination with flood fill algorithm, we propose a discrete noise removal algorithm. In standardizing modules, combining snake segmentation with vertical segmentation, effectively segment the tilting characters[4]. After relocation, we standardize characters using the bilinear interpolation algorithm. Based on Google’s open source library of character recognition, tess-two, a branch of tesseract, we identify the word.We provide local and online translate. The local translation uses android SQLite database. Query the.db file on the SD card. The word library is small and the explanation is rough. The online translation uses Youdao translation API. The word library is large and the explanation is detailed.The experimental results show that even under the dim light and a small amount of noise, the system also has very good recognition rate, reaching more than 90% accuracy.