The Design And Implementation Of Optical Character Recognition Application Based On Android Platform

With the development of the digital age, the amount of information have experienced explosive economic growth in recent years. However, the speed of inputing imformation into digital divices is far less than the speed of the information growing. Optical character recognition (OCR) can detect and identify information automatically, effectively solve the problem of information entry speed and accuracy rate. Currently, optical character recognition based on PC devices have been widely used in office automation and industrial intelligence, etc. However, due to the weak portability, it is difficult to process information in the daily life. With the development of hardware of smartphone and the popularity of the Android operation system recently, developing optical character recognition application on smartphone is possible, so, how to develop an effective application has become an urgent problem.This paper studied the priori knowledge and the basic theory of optical character recognition on PC, then applied on mobile devices, designed and implemented an optical character recognition application based on Android platform. This paper focues on the function of recognition and the interface design.The main work of this paper include:1) Requirements analysis and preliminary design of the application, divide the application into three modules:image acquisition module, optical charactor recognition module and information storage module. Image acquisition module is responsible for collecting images containing character information; Character recognition module is responsible for processing the acquired images, identifing and outputing the recognition result; Information storage module is responsible for storing the original image, the recognition result and editing operations on them.2) Detailed design and implementation of applications. The image acquisition module has two ways to capture images, calls the interface of camera and photo albums respectively. For optical charactor recognition module, combining research and analysis of the literature of the usage scenarios, designed a series of optical character recognition algorithm. Because of the complexity of different usage scenarios, using gray, smoothing, binarization, morphological operations and other image pre-processing algorithms to enhance the images, highlight the difference between the backgrounds and identify targets. Then, figure out the image of the character region, character segmentation and character normalized, put the normalized image to the SVM model to recognize.Then we get the recognition result. For information storage module, using an embedded database engine named SQLite to store the recognition result as data files in the SD card.3) Test and evaluation of the application. Testing the application in the actual use cases, analyzing the result by precision rate, recall rate and F score.This paper aims at using the compute power increasingly of the mobile platform, implementing an optical character recognition application which can be used in daily life, allowing users to get and store information required.