隨著科技日新月異的進步，
掃瞄儀器 越來越大眾化的情況下，我們可以看到
越來越多的圖像辨識技術以及光學字元辨識之技術 漸漸地 與我們的生活密不可
分。光學字元辨識， 一種文字自動輸入的方法， 顧名思義就是以手機或者相機等
電子設備拍 下所需要辨識的辨識之圖片後，透過偵測 光線 亮與暗的方式來決定字
的外型，再利用光學字元辨識的技術將其轉換為需要辨識之文字的 過程。 可 簡易
分成三個部分 : 1.影像掃描器 2.光學字元辨識軟體 3.輸出介面，先透過影像掃描
器將輸入之影像掃描過後 並且將圖片有損毀的部分先做影像前處理 接著 將其 處
理結果 輸入光學字元辨識軟體辨識，再將辨識結果轉換成特定文件類型 (PDF
WORD…等 再 轉交由輸出介面顯示出結果或者儲存到特定位置 。 本論文為了
有效的增加辨識率， 結合影像辨識及光學字元技術 主要為提高中文辨識率，透
過實驗的結果可以看出本論文所提出之研究方法可以有效的識別影像中的中文
字。
關鍵詞：
影像處理 、光學字元識別、物體偵測。

With the develop of the technology and the scanning instrument becoming more popularization. We know the Image Recognition Technology and Optical Character Recognition inseparable from our lives gradually. Optical Character Recognition is a method of automatic input word, after we use the cellphone or camera to take the pictures that needed to recognize, Determine the shape of the word by detecting light and dark, and the process if using Optical Character Recongnition technology to convert it into word that needed to be recognized. We can simply divided into three parts :1.Image scanner 2. Optical Character Recognition software 3. Output interface , After scanning the input image through an image scanner and pre-processing the damaged part of the image, then input the processing result into the Optical Character Recognition software for recognition , and then convert the recognition result into a specifi file type(PDF , WORD…).and then forward the result to the output interface or save it to a specific location.In order to effectively increase the recognition rate,this paper combines image recognition and Optical Character Recognition technology, mainly to improve the Chinese recognition rate.Through the experimental results, it can be seen that the research method proposed in this paper can effectively recognize Chinese characters in images.

摘要 i
Abstract ii
誌謝 iii
目錄 v
表目錄 vii
圖目錄 viii
1、緒論 1
1.1研究背景 1
1.2研究動機與目的 2
1.3論文架構 3
2、文獻回顧 4
2.1光學字元識別 4
2.2影像前處理 5
2.2.1灰階化 6
2.2.2平均濾波 6
2.2.3邊緣偵測[11][16] 7
2.2.1二值化[7] 9
2.2.2梯度運算 11
3、研究方法 12
3.1本實驗之系統流程圖介紹 13
3.1.1選取圖像、路徑、以及數入圖像 13
3.1.2影像灰階化、平均濾波、利用LAPLACIAN尋找邊緣 14
3.1.3尋找字體輪廓、圈選ROI 14
3.1.4進行文字識別、輸出 16
3.2廣告圖之字元識別系統 17
3.2.1影像前處理技術 17
3.2.2梯度運算 21
3.2.3尋找輪廓，圈選ROI[20] 22
3.2.4字元識別 23
4、研究結果與分析 24
4.1系統之前置測試的研究結果與分析 24
4.2系統的研究結果 33
5、結論與未來展望 35
6、參考文獻 36

[1]M. Abdalla; B. Nagy (2018, April). Dilation and Erosion on the Triangular Tessellation: An Independent Approach. In IEEE Access (Volume: 6 ) .IEEE.
[2]L. Anisa; S. Nina; A.N. Raden; W.S. Valentina; P.C. Setyo; D.P. Stephani (2019, September). Google vs. Instagram Machine Translation: Multilingual Application Program Interface Errors in Translating Procedure Text Genre. In 2019 International Seminar on Application for Technology of Information and Communication (iSemantic). IEEE
[3]G. Cheung; E. Magli; Y. Tanaka; M. K. Ng (2018, April). Graph Spectral Image Processing. In Proceedings of the IEEE ( Volume: 106 , Issue: 5 , 2018,May). IEEE
[4]B. Chen; L. Cao (2017, December), An Optimized Algorithm for Calculating the Average Path Length of Complex Network. In 2017 10th International Symposium on Computational Intelligence and Design (ISCID). (pp. 334-337). IEEE.
[5]E. Genc; H.R. Shin; J.S. Park (2018, October). Number Recognition of Parts Book Schematics using Convolutional Recurrent Neural Network. In 2018 18th International Conference on Control, Automation and Systems (ICCAS). (pp1190-1192). IEEE.
[6]J. Guan; X. Ye; J. Gao; Q. Liu (2009, November). The Software Communication Architecture specification: Evolution and trends. In 2009 Asia-Pacific Conference on Computational Intelligence and Industrial Applications (PACIIA). (pp. 341-344). IEEE.
[7]X. Hu; S. Wu; W. Xu (2019, June). Adaptive Binarization for Degraded Document Image via Contrast Enhancement. In 2019 14th IEEE Conference on Industrial Electronics and Applications (ICIEA). (pp. 2363-2667). IEEE.
[8]L. Jin; E. Song; L. Li; X. Li (2013, September). A quaternion gradient operator for color image edge detection. In 2013 IEEE International Conference on Image Processing (pp. 3040-3044). IEEE.
[9]M. Mahmut ; Y. Genc(2019, December). A Deep-Learning Approach to Optical Character Recognition for Uighur Language.In 2019 International Conference on Advances in Computing, Communication and Control (ICAC3) (pp 1-6). IEEE.
[10]H. Mehta; S. Singla; A. Mahajan (2016, May). Optical character recognition (OCR) system for Roman script & English language using Artificial Neural Network (ANN) classifier. In 2016 International Conference on Research Advances in Integrated Navigation Systems (RAINS). (pp1-5). IEEE.
[11]R. Minetto.; N. Thome; M. Cord; J. Stolfi; F. Precioso.; J. Guyomard,; & N.J. Leite. (2011, November). Text detection and recognition in urban scenes. In 2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops) (pp. 227-234). IEEE.
[12]W.A. Mustafa; H. Yazid; S.B. Yaacob (2014, November). lllumination normalization of non-uniform images based on double mean filtering. In 2014 IEEE International Conference on Control System, Computing and Engineering (ICCSCE 2014) (pp. 366-371). IEEE.
[13]S. Ghada; E.H. Elsayed; R. Mohamed (2018, November). A Generic OCR Using Deep Siamese Convolution Neural Networks. In 2018 IEEE 9th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON). (pp1238-1244). IEEE.
[14]Y.M. Su; H.W. Peng; K.W. Huang; C.S. Yang (2019, November). Image processing technology for text recognition. In 2019 International Conference on Technologies and Applications of Artificial Intelligence (TAAI). (pp1-5). IEEE.
[15]P. Sun ; Q. Xie ; Z. Wu ; X. Feng ; J. Cai ; Y. Jiang (2019,October).Yi Characters Recognition Based on Tesseract-OCR. In 2019 IEEE 3rd Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC) (pp.102-106). IEEE.
[16]S. Suwanmanee; S. Chatpun; P. Cabrales (2013, October). Comparison of video image edge detection operators on red blood cells in microvasculature. In The 6th 2013 Biomedical Engineering International Conference. IEEE.
[17]K.H. Talukder; T. Mallick (2014, December). Connected component based approach for text extraction from color image.In 2014 17th International Conference on Computer and Information Technology (ICCIT). (pp. 204-209). IEEE.
[18]Y.C. Xing; R. Li; L. Cheng; Z. Wu (2018, December). Research on Curved Chinese Document Correction Based on Deep Neural Network. In 2018 11th International Symposium on Computational Intelligence and Design (ISCID). (pp342-345). IEEE.
[19]H.M. Ye; S.L. Yan; X. Bai (2017, April). Application of Switching Median Filter in Two-Dimensional Otsu Image Segmentation. 2017 International Conference on Network and Information Systems for Computers (ICNISC) (html). IEEE.
[20]V. Zilvan; F.N. Azizah (2018, November). Region of Interest Mining Using Stay Point Detection and Point Region Quadtree. In 2018 5th International Conference on Data and Software Engineering (ICoDSE). (pp. 1-5). IEEE.