透過您的圖書館登入
IP:18.222.184.162
  • 學位論文

應用手機及遠端系統檢測市售白米等級的初步建立

Preliminary Establishment of Application Smart Phone and Remote System to Detect Quality Degree of Polished Rice Obtained from Market

指導教授 : 謝清祿

摘要


摘要 學號:M10344009 論文題目:應用手機及遠端系統檢測市售白米等級系統初步建立 總頁數:69頁 學校名稱:國立屏東科技大學 系別:生物機電工程系 畢業年月:2016年7月 學位別:碩士學位 研究生:陳弘晸 指導教授:謝清祿 博士 論文摘要內容: 白米是目前台灣人最常食用的主食之一,但偶有不肖業者添加劣質米的白米流入市場。同時民眾對於智慧型手機的使用率日益漸高,其中擁有行動網路的使用者也是逐年攀升。本研究設定兩個方案來實現,方案一是將白米正常粒與異常粒判別方程式編寫成手機的APP,讓使用者可以直接使用手機拍照,並且利用APP做立即的白米影像分析快速得到結果,而且不需要網路就可以進行分析,但是影像處理的龐大資料對手機來說十分費力,改為開發方案二透過手機APP超連結讓使用者在有網路的地方都可以快速連結到網頁,隨時隨地上傳白米影像進行分析檢測。上傳的影像將會由伺服器端透過影像處理的方式將白米正常粒、碎粒、白粉質粒、黃粒、色斑粒、損害粒及異型粒,將每一粒白米分類出來後由CNS的白米分級標準進行分級,以方便判斷出使用者白米影像的等級,分析結束後會以電子郵件的方式將檢測的結果發送至使用者的信箱,以供使用者的參考。方案一分析了白粉質粒、黃粒、色斑粒三種異常粒,準確率分別為81.1%、79.4%、36.2%。在方案二中各種白米異常粒的判別中,以碎粒的判別準確率最高達到95.3%,損害粒的準確率最低為47.7%,白粉質粒準確率為86.1%,黃粒準確率為83.9%,色斑粒準確粒為53.2%,異型粒準確率為78.1%。 關鍵詞:白米、影像處理、手機應用程式、遠端系統

並列摘要


Abstract Student ID:M10344009 Title of Thesis : Preliminary Establishment of Application Smart Phone and Remote System to Detect Quality Degree of Polished Rice Obtained from Market Total pages:69pages Name of Institute:National Pingtung University of Science and Technology Name of Department:Department of Biomechatronics Engineering Date of Graduation:July, 2015 Degree Conferred:Master Name of Student:Hong-Jhen Chen Adviser:Dr. Ching-Lu Hsieh The Contents of Abstract in this Thesis: Rice is the staple food in Taiwan, but occasionally some unqualified rice was found in market. Nowadays mobile phone or smart phone becomes popular for modern life as well as the mobile internet services. Thus, this study developed two solutions to detect the quality of rice by smart phone APP (application). Solution one developed an Android APP that can take image of rice sample and analyze the rice quality. This approach does not need Internet connection. But it cost a lot of mobile phone memory space. Therefore, the second resolution was proposed which uploaded the rice image to web server via Internet. The server equips image analysis program that can automatically analyze the uploaded image and evaluate the rice quality based on the proportion of broken kernel, chalky kernel, yellow kernel, color-streaked kernel, damaged kernel, and off-type kernel. Their quality level on the criteria of CNS will be sent by Email to the users. The evaluation accuracy of solution one reached 81.1%, 79.4%, and 36.2% for chalky kernel, yellow kernel, and color-streaked kernel. Solution two has highest evaluation accuracy of 95.3% for broken kernel while the lowest of 47.7% for damaged kernel. The evaluation accuracy for chalky kernel, yellow kernel, color-streaked kernel, and off-type kernel accuracy were 86.1%, 83.9%, 53.2% and 78.1% respectively. Keywords:rice, image processing, APP, Remote system

並列關鍵字

rice image processing APP Remote system

參考文獻


林連雄,2007,近紅外線影像及發光二極體檢測系統偵測白米內部品質之研究,博士論文,臺灣大學,生物產業機電工程學研究所。
萬一怒、溫惠雯,1998,「糙米品質自動檢測分級系統– (二)糙米外觀品質檢測之研究」,農業機械學刊7(2),第13-27頁。
萬一怒、林建銘、廖家興、周瑞仁,1997,「胴裂米反光特性的分析及應用」,農業機械學刊6(3),第89-99頁。
曾嘉邵,2010,新型白米篩選機構之開發,碩士論文,大同大學,機械工程學系(所)。
劉小瑄,2011,新型白米外觀品質檢測系統之研製,碩士論文,中興大學,生物產業機電工程學系所。

延伸閱讀