透過您的圖書館登入
IP:3.141.2.96
  • 期刊
  • OpenAccess

砷化銦鎵多通道檢波器水果檢測系統之研製

Development of an InGaAs Multichannel Detector Inspecting System for Fruit

摘要


目前市售水果近紅外線線上檢測機均使用矽(Si)質檢波器,該材質受限於檢測波長之關係,無法檢測高於1100nm之波長,文獻記載水果檢測低於1100nm之波長適用於糖度,高於1100nm之波長則適用於酸度。砷化銦鎵(InGaAs)檢波器之檢測波長在800~1700nm,此段波長除可應用於糖度外,依據文獻未來在酸度的應用性更佳,但砷化銦鎵檢波器具高暗電流,以及光子能量會隨檢測波長增加而變弱等特性,又衍生出其他問題,致使檢測效果受到考驗。本研究以砷化銦鎵檢波器,配合照明光源、聚光鏡頭、單光儀及其控制系統,研製一套以900~1600nm波長為範圍之線上型近紅外線水果檢測系統。用光學元件設計系統之照明光源及聚光鏡頭,以加強系統之檢測信號。經試驗發現檢波器溫度-10℃、曝光時間200ms、增益值5、A/D轉換12bits為系統操作之最佳條件。另使用尼龍板簡化傳統波長校正方法,建立校正模式之結果為R^2=0.999及SEC=1.546。為瞭解研製系統之檢測效果,量測100顆印度棗之吸收光譜,建立預測糖度之最佳檢量線,得到R^2、SECV、1-VR及SEC值各為0.8349、0.9392、0.6017及0.6000之結果,可證明本套線上型近紅外線水果檢測系統具有高度準確性之實用價值,能於水果分級上做即時之內部品質檢測。

並列摘要


Si photodiode array is widely adapted in the commercial fruit online inspecting systems of near infrared (NIR). But the effective scope on detecting wavelength of a Si-type detector prevented it from measuring over 1100nm wavelength. Reports indicated the best wavelength of utilizing NIR to inspect fruit was below 1100nm for sugar content; and above 1100nm for acidity. The detecting wavelength of 800~l700nm was effective working wavelength range for an InGaAs detector, which can be used not only to predict the sugar content but also to detect acidity for fruits in future. However, the characters of a high dark current and the photo-energy getting weak along the wavelength of an InGaAs detector induced other problems. This is also inhibited the evaluation function of an InGaAs detector. In this research, an InGaAs multichannel detector in the range of 900~1600nm wavelength was chosen to develop a NIR fruit online inspecting system with two light sources, a collecting lens, a monochrometor, and a control unit. The light source and collecting lens of system made of optical components were designed to enhance the detecting signal. The results of tests showed that detector temperature at -10℃, exposure time of 200ms, gain value of 5, and 12 bits A/D were the optimum parameters for operation A method for wavelength calibration with nylon board of R^2=0.999 and SEC=1.546 was established to simplify the work of traditional way. The performance of the developed system was validated by measuring absorption spectrum of 100 Indian jujubes. The results were used to establish the effective calibration model to predict the sugar content of Indian jujubes, the values of R^2, SECV, 1-VR and SEC are 0.8349, 0.9392, 0.6017, and 0.6000, respectively. These results indicated that the developed NIR inspecting system can be used to assess the internal quality evaluation of a fruit grading system in real time.

被引用紀錄


鄭宇帆(2009)。高光譜影像於龍膽指標成份之檢測應用〔碩士論文,國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342/NTU.2009.10107

延伸閱讀