臺灣博碩士論文加值系統

本論文延續實驗室前兩代學長以條碼影像辨識為基礎的圖書盤點系統研究，特別著重在建立條碼影像即時傳送以及後端處理結果即時回傳至盤點現場的作業流程。實驗室先前兩代的研究重點在針對盤點現場拍攝的照片中大量的條碼影像進行定位、分割與辨識，而在條碼辨識後仍有部分工作需回到現場處理，所以整個盤點流程大致分為三個階段。本論文研究的目的是要將盤點流程簡化為一階段，因此工作的重點是支援影像即時傳輸與辨識結果即時回傳至盤點現場的軟體開發與系統整合。
本論文開發Android系統上的應用程式，以智慧型行動裝置擷取並傳送圖書影像，並整合伺服器端的辨識與比對軟體系統。使用流程為啟動裝置上的相機拍照後傳送至後端，照片進入伺服器後進行影像處理、條碼分析以及資料庫比對的作業，再將結果條列回傳至前端的行動裝置，盤點人員依據回傳結果處理書架上多出的書本或確認遺失狀況。
我們以三支不同廠牌或型號的智慧型手機進行實驗，每支手機拍攝300張實驗照片。統計每支手機所拍照片的辨識結果得知，每張照片可正確辨識的條碼數達七成（含）以上的比率分別為49.0%、82.3%及85.0%。辨識成功率的高低取決於成像品質，而與規格所列的照相鏡頭解析度無明顯關聯。在拍攝條件及成像品質優良的情況下，本系統藉由智慧型行動裝置與後端伺服器的協同運作有效輔助圖書盤點作業，提升盤點人員整體工作效率。

This thesis is a continuation of an effort to develop a barcode image based library inventory system. This research represents the third phase of the development in progress which emphasizes on establishing a highly efficient workflow, where photo images containing many barcodes are sent on-site through the Wi-Fi network to the server for processing and the results of database matching and related instructions are sent right back to the personnel inventorying. The development effort of this phase aims at transforming the original three-step inventory workflow into a near-instant one-step process. Thus, our work has concentrated on developing and integrating software applications for both mobile devices at the front-end and computer at the back-end.
To streamline the inventory process, we have developed programs for Android mobile devices to send pictures of books on the shelves and to receive and display the corresponding instructions from the back-end server. Software components on the server side were also developed and integrated, so that the inventory personnel can manually handle the misplaced books on-site as well as confirm whether a book is truly missing from the shelf it belongs to.
To evaluate the effectiveness and performance of the inventory system, we have employed three smartphones of different brands or models, each taking a group of 300 pictures for our experiment. Each picture taken contains from twenty-something to around fifty barcodes. If at least 70% of the barcodes in one picture can be correctly extracted is regarded as a successful recognition, then the ratios of successful recognition of the three picture groups are 49.0%, 82.3%, and 85.0%, respectively. The results indicate that how well the barcodes can be recognized is largely dictated by overall quality of the picture and has much less to do with camera resolution stated in the specification of a smartphone. Assuming proper lighting conditions and with mobile devices capable of producing high-quality pictures, our one-step library inventory system has a great potential to save time and manpower for libraries running with limited staff.

致謝 I
摘要 II
Abstract III
目錄 IV
表目錄 V
圖目錄 VI
第一章 緒論 1
1.1 研究動機 1
1.2 研究目的與方法 1
1.3 論文內容提要 3
第二章 前置處理與條碼定位 4
2.1 影像灰階 4
2.2 直方圖均衡化（Histogram Equalization） 5
2.3 中值濾波（Median Filter） 7
2.4 光變化量定位法 9
2.5 二值化（Thresholding） 11
2.5.1 全域性二值化 11
2.5.2 區域性二值化 13
2.6 物件連通 15
第三章 系統實作 21
3.1 系統組成 21
3.2 系統運作流程 22
3.3 設計細節 24
3.3.1 開啟Android的使用權限 24
3.3.2 使用Android拍照功能 24
3.3.3 HTTP POST 26
3.3.4 非同步任務 (AsyncTask) 27
第四章 實驗結果 29
4.1 實驗環境 29
4.2 實驗設備 29
4.3 實驗結果 30
第五章 結論與未來展望 34
參考文獻 35

[1] 林柏亨，圖書盤點系統之條碼影像前置處理演算法優化與效能提升，國立台灣海洋大學電機工程學系碩士學位論文，2013。
[2] Wikipedia, BMP <https://zh.wikipedia.org/wiki/BMP>,January 2016.
[3] Wikipedia, JPEG<https://zh.wikipedia.org/wiki/JPEG>, January 2016.
[4] 謬紹剛，數位影像處理，普林斯頓國際有限公司，2009。
[5] Wikipedia, Histogram Equalization<https://en.wikipedia.org/wiki/Histogram_equalization>, January 2016.
[6] Pavel Šimurda, “Barcode Localization in Image,” Proceedings of the 17th Conference STUDENT EEICT 2011 ,Vol 1, pp.169–171, Brno, Czech Republic, 2011.
[7] N. Otsu, “A Thresholding Selection Method from Gray Level Histogram,” IEEE Transactions on System, Man and Cybernetics, Vol. 9, No. 1, pp 62–66, 1979.
[8] J. Sauvola, T. Seppänen, S. Haapakoski, and M. Pietikiiinen, “Adaptive Document Binarization,” Proceedings of IEEE Conference on Document Analysis and Recognition, pp. 147–152, 1997.
[9] Google, Manifest.permission,<http://developer.android.com/reference/android/Manifest.permission.html>, May 2015.
[10] Google, Android Developers<http://developer.android.com/training/camera/photobasics.html>, May 2015.
[11] Google, Intent<http://developer.android.com/reference/android/content/Intent.html>, May 2015.
[12] RFC 7231 Hypertext Transfer Protocol (HTTP/1.1): Semantics and Content, <http://tools.ietf.org/html/rfc7231#section-4.3.3>, October 2015.
[13] Anders Göransson，Android 高效能多執行緒（楊仁和譯），碁峯資訊股份有限公司，2015。
[14] Google, AsyncTask<http://developer.android.com/reference/android/os/AsyncTask.html>, October 2015.
[15] Ravi Tamada, Android Uploading Camera Image, Video to Server with Progress Bar<http://www.androidhive.info/2014/12/android-uploading-camera-image-video-to-server-with-progress-bar/>, October 2015.