针对开源识别引擎Tesseract-OCR对噪点多、亮度不均匀及规格不统一的图像识别效果不佳的情况,设计和实现了一种基于Android平台,能大幅度提高质量不高图像识别率的图文同步识别系统.实现了预览同步识别、联网上传识别、图像批量识别等功能,通过对图像进行消噪、亮度均衡及阈值分割等质量增强算法处理,提高了图像的最终识别率.新颖的同步识别模式有别于传统的图文识别软件,使用户在预览图像时能够即时看到识别效果,给使用者带来一种全新的用户体验.
Tesseract-OCR,an open source OCR engine,has less efficient with the images which have much noise,uneven brightness,and various specifications.This paper designs and implements a synchronized OCR system which can improve recognition rates with low-quality images significantly based on the android-based phone.This system not only implements the preview &synchronized OCR function,network & upload OCR function and batch of images OCR function,but also improves final recognition rates by the way of image denoising,balancing brightness and thresholding segmentation.Synchronized OCR as an innovative model is different from the traditional OCR software,users can see the OCR results immediately when they see the preview images,and it will bring a totally new experience to users.