臺灣博碩士論文加值系統

本論文為歷經前三代實驗室學長以條碼影像辨識為基礎的圖書盤點系統研究之延續。第一代及第二代的研究主要是利用數位相機在盤點現場拍攝大量書脊貼有條碼之藏書的照片，再將這些照片輸入到後端PC上進行條碼影像定位、分割以及辨識處理進而得到分析結果。第三代的研究則是著重在利用智慧型行動裝置進行拍照並將影像即時傳輸至後端PC，大幅簡化圖書盤點流程。本論文研究基於智慧型行動裝置性能不斷提升的事實，除延續第三代利用智慧型行動裝置拍照外，更直接在智慧型行動裝置上進行大量條碼影像分析與辨識的運算，並且將辨識結果和遠端資料庫比對後，即時以顏色標記顯示在現場盤點照片上。此處理流程除了大幅降低盤點系統的硬體建置成本外，對於盤點工作流程的簡化極為顯著。此外，本研究也適當縮小黏貼於書脊上的條碼，減少相鄰條碼照成分析困難的機會。
本論文軟體系統實現於Android行動裝置平台上，以智慧型行動裝置內建的照相功能擷取圖書影像並在行動裝置上完成條碼辨識的工作，再搭配遠端資料庫以純文字的方式傳輸資料完成比對工作，盤點人員只需依據比對的結果處理書架上不應出現或遺漏的藏書。
在效能驗證方面，我們以三支不同廠牌與規格的智慧型手機，分別針對三種不同尺寸之條碼進行測試，每支智慧型手機對不同尺寸之條碼各拍攝30張相片(共計270張測試樣本)並統計其辨識率。我們的實驗結果顯示，成像品質(解析度)對於辨識率的高低有著決定性的影響；再者，條碼的尺寸與辨識率並非正相關，在一般的拍攝條件下，條碼尺寸越大反而會降低其辨識率。相較於前一代圖書盤點系統，在兼顧盤點效率以及縮小條碼尺寸之前提下，本論文已能將辨識率提高至九成。

This thesis is the fourth generation of a continuing research effort to build a barcode image based library inventory system. With the first two generations of the system, digital cameras were used to take pictures of books on the shelves, mainly for capturing the barcode labels stuck to the spines of the books. Those pictures then went through a series of processing steps done on a backend server, including barcode localization, segmentation, and identification. The third generation of the system emphasized the use of smart mobile devices for capturing and transmitting the book images wirelessly to the backend server and receiving the results of analyses sent back by the server. The use of smart mobile devices effectively improved the efficiency of the inventory process.
The current thesis research recognizes the advances of smart mobile technology and decides to take full advantage of the processing power available on modern smart phones. By doing the analysis and recognition of multiple barcodes directly on the very same smart phone which also takes the pictures, we are able to reduce the hardware cost of the inventory system significantly. Furthermore, after comparing the data extracted from the barcode images with the records in the library database, color-coded marks will appear on the original pictures to show which books are correctly shelfed and which are misplaced. Books missing from the shelf are also listed instantly. Through this process, the workload of inventory personnel is reduced enormously.
The current library inventory system has been implemented as an App on the Android system. To evaluate the system’s performance, we use three smart phones from different vendors for picture taking and recognizing the barcodes of three different sizes. Each smart phone was arranged to take three groups of pictures with each group consisting of 30 pictures. Within each group, all the books appearing in the picture were stuck with barcode labels of one particular size. A total of 270 test pictures were taken and analyzed and, from which, the average recognition rates of each smart phone against each size of barcode were calculated. The results of our experiment indicate that picture quality, dictated by the resolution, has the decisive impact on the level of recognition rate. However, the average recognition rate is counterintuitively in inverse proportion to the size of the barcode. With proper choice of the barcode size, the overall barcode recognition rate can reach over 90% by the current implementation. When compared to the previous generations, the current generation is easier to use, perform faster, and cheaper to set up.

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 論文內容概覽 ．．．．．．．．．．2
第二章 一維條碼概述．．．．．．．．．3
2.1 一維條碼簡介 ．．．．．．．．．．3
2.2 一維條碼讀取原理．．．．．．．．．5
2.3 128碼之編碼與解碼．．．．．．．．6
第三章 文獻回顧．．．．．．．．．．．10
3.1 條碼定位．．．．．．．．．．．．10
第四章 系統實作．．．．．．．．．．17
4.1 系統組成與運作流程．．．．．．．17
4.2 設計細節．．．．．．．．．．．．．19
4.2.1 開啟Android的使用權限．．．．．19
4.2.2 啟用Android相機．．．．．．．．20
4.2.3 使用條碼偵測API．．．．．．．．20
4.2.4 HTTP GET．．．．．．．．．．．21
4.2.5 如何標記條碼辨識後之書籍．．．．23
第五章 實驗結果．．．．．．．．．．．．．25
5.1 實驗環境．．．．．．．．．．．．．．25
5.2 實驗設備．．．．．．．．．．．．．．28
5.3 實驗結果．．．．．．．．．．．．．．28
5.3.1 條碼辨識率之實驗結果．．．．．．．29
5.3.2 盤點速度實測之實驗結果．．．．．．34
5.3.3 即時標記功能與偵測遺漏書籍之實驗結果 ．．．．35
第六章 結論與未來展望．．．．．．．．．．41
6.1 結論 ．．．．．．．．．．．．．．41
6.2 未來展望．．．．．．．．．．．．．．41
參考文獻 ．．．．．．．．．．．．．．43
附錄 ．．．．．．．．．．．．．．45

參考文獻
[1] Pavel Šimurda, “Barcode Localization in Image,” Proceedings of the 17th Conference STUDENT EEICT 2011, Vol. 1, pp. 169–171, Brno, Czech Republic, 2011.
[2] A. K. Jain and Y. Chen, “Bar code Localization using texture analysis,” IEEE Conference on Document Analysis and Recognition, pp. 41–44, 1993
[3] 鐘子毅，基於條碼辨識的圖書館藏書盤點系統，國立台灣海洋大學電機工程學系碩士學位論文, 2013。
[4] 朱偉傑，在嵌入式系統上實線之不完整條碼辨識系統，台北科技大學電機工程系碩士學位論文, 2009。
[5] 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。
[6] 吳明衡，以智慧型裝置實現的條碼影像及時圖書盤點系統，國立台灣海洋大學電機工程學系碩士學位論文，2016。
[7] F. V. Reischach, S. Karpischek, R. Adelmann, and F. Michahelles, “Evaluation of 1D barcode scanning on mobile phones,” Internet of Things 2010 Conference (IoT2010), 2010.
[8] 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.
[9] 陳柏珽，分而治之演算法之探討，逢甲大學應用數學研究所碩士論文，2011。
[10] Google, Track Face and Barcode <https://developers.google.com/vision/multi-tracker-tutorial> , July 2016.
[11] Google, Manifest.permission <http://developer.android.com/reference/
android/Manifest.permission.html>, July 2016.
[12] Google, Android Developers <https://developer.android.com/training/building-multimedia.html >, July 2016.
[13] Google, Android Developers <https://developer.android.com/reference/android/
graphics/BitmapFactory.html> , July 2016.
[14] w3schools, HTML < http://www.w3schools.com/tags/ref_httpmethods.asp >, July 2016.
[15] Wikipedia, JSON <https://zh.wikipedia.org/wiki/JSON>, July 2016.
[16] Google, Volley <https://developer.android.com/training/volley/index.html>, July 2016.
[17] Google User Guide, Gson <https://sites.google.com/site/gson/gson-user-guide>, July 2016.
[18] Google, Canvas<https://developer.android.com/reference/android/graphics/ Canvas.html >, July 2016.
[19] 圖片來源 < http://www.barcodelabelhk.com/education/how-to-scan-barcode.htm>, November 2016.
[20] 圖片來源 < http://www.appsbarcode.com/code%20128.php>, November 2016.
[21] 圖片來源 <https://magiclen.org/easy-barcode-scanner/>, November 2016.
[22] 圖片來源 < http://jamesyvi67patton.pixnet.net/blog/post/31211042-keurig-k-cup%E5%96%AE%E6%9D%AF%E5%92%96%E5%95%A1%E6%A9%9F%2Biphone-app-shopsavvy>, Novebmer 2016.