基于Android平臺文字識別應用的設計與實現
發(fā)布時間：2018-08-16 20:00
【摘要】：隨著OCR技術的日漸成熟,移動互聯(lián)網的蓬勃發(fā)展,擁有OCR技術的移動終端產品逐漸深入到了人們的日常生活。移動終端的便攜性使其適合作為大規(guī)模移動OCR商用平臺。本文對Android框架和JNI技術進行綜述,將Android平臺與OCR技術結合,實現了基于Android平臺文字識別系統(tǒng)應用的設計與開發(fā)。論文主要工作內容有:(1)針對文字識別系統(tǒng)相應的技術和算法進行了概括。研究對自然場景下所拍攝圖像進行文本標記的算法,為后續(xù)OCR部分中的文字分割提供相關參數。同時研究圖像預處理相關算法,在極值中值濾波基礎上,提出了一種適用于Android平臺的圖像預處理方案。(2)分析了 Tesseract工作流程與相關算法,利用JNI進行Android應用程序和Tesseract庫的交互,并將其移植到Android平臺,作為移動終端文字識別的引擎。設計了系統(tǒng)的框架和實現方式。最終實現了對英文字符和中文字符的識別。(3)測試和分析了系統(tǒng)性能。根據分析結果驗證了 OCR系統(tǒng)性能良好,并針對不足之處給出了有效的改進方案,為進一步深入研究提供了可靠的數據支持。本文實現了 Android終端上中英文字符識別應用的設計、開發(fā),對OCR理論研究和商業(yè)應用都十分有意義。
[Abstract]:With the maturity of OCR technology and the vigorous development of mobile Internet, mobile terminal products with OCR technology have gradually penetrated into people's daily life. The portability of mobile terminal makes it suitable for large scale mobile OCR commercial platform. In this paper, Android framework and JNI technology are summarized, and the design and development of character recognition system based on Android platform are realized by combining Android platform with OCR technology. The main contents of this paper are as follows: (1) the corresponding techniques and algorithms of character recognition system are summarized. This paper studies the text marking algorithm of the images taken in natural scene, and provides the relevant parameters for the text segmentation in the following OCR part. At the same time, the correlation algorithm of image preprocessing is studied. On the basis of extremum median filter, an image preprocessing scheme suitable for Android platform is proposed. (2) the Tesseract workflow and related algorithms are analyzed, and the interaction between Android application program and Tesseract library is carried out by using JNI. And transplant it to Android platform, as mobile terminal text recognition engine. The framework and implementation of the system are designed. Finally, the recognition of English characters and Chinese characters is realized. (3) the system performance is tested and analyzed. According to the analysis results, the performance of the OCR system is proved to be good, and an effective improvement scheme is given in view of the shortcomings, which provides reliable data support for further research. This paper implements the design and development of Chinese and English character recognition application on Android terminal, which is of great significance to the research of OCR theory and commercial application.
【學位授予單位】：北京郵電大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：TP391.41;TP316
[Abstract]:With the maturity of OCR technology and the vigorous development of mobile Internet, mobile terminal products with OCR technology have gradually penetrated into people's daily life. The portability of mobile terminal makes it suitable for large scale mobile OCR commercial platform. In this paper, Android framework and JNI technology are summarized, and the design and development of character recognition system based on Android platform are realized by combining Android platform with OCR technology. The main contents of this paper are as follows: (1) the corresponding techniques and algorithms of character recognition system are summarized. This paper studies the text marking algorithm of the images taken in natural scene, and provides the relevant parameters for the text segmentation in the following OCR part. At the same time, the correlation algorithm of image preprocessing is studied. On the basis of extremum median filter, an image preprocessing scheme suitable for Android platform is proposed. (2) the Tesseract workflow and related algorithms are analyzed, and the interaction between Android application program and Tesseract library is carried out by using JNI. And transplant it to Android platform, as mobile terminal text recognition engine. The framework and implementation of the system are designed. Finally, the recognition of English characters and Chinese characters is realized. (3) the system performance is tested and analyzed. According to the analysis results, the performance of the OCR system is proved to be good, and an effective improvement scheme is given in view of the shortcomings, which provides reliable data support for further research. This paper implements the design and development of Chinese and English character recognition application on Android terminal, which is of great significance to the research of OCR theory and commercial application.
【學位授予單位】：北京郵電大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：TP391.41;TP316
【參考文獻】
相關期刊論文 前10條
1 劉淼;楊鎮(zhèn)豪;謝韻玲;謝冬青;唐春明;;Android圖文同步識別系統(tǒng)的設計和實現[J];計算機工程與設計;2014年06期
2 童立靖;張艷;舒巍;占國亮;錢W，
本文編號：2187031
本文鏈接：http://sikaile.net/kejilunwen/ruanjiangongchenglunwen/2187031.html
