以螢光影像技術檢測蘋果之褐變

Abstract

水果品質檢測是農業發展上一大課題，水果經過完善品質檢測後，除了能夠保障消費者的權益，也能夠提高農民的收益。褐變是水果內部產生化學變化使果皮或果肉顏色產生黑褐色改變的現象；造成水果褐變的原因有很多，如儲存於不當之氣調環境下、運輸過程中發生碰撞等，上述兩者皆是因為水果中的多酚氧化酶polyphenoloxidase（PPO）催化酚類化合物使其氧化並聚合成不溶的深咖啡色聚合物美拉寧（melanins）而造成，如果能以非破壞的檢測方法篩選內部已褐變但外觀尚未有明顯變化的水果，即能確保水果之品質。葉綠素螢光量測為一種非破壞性的檢測方法，早期大都採用單點量測，對象物以植物葉片居多，近期則發展二維影像的多光譜螢光影像系統，可應用於生物材料之內部品質檢測。水果淺層之撞擊型褐變，對表層的葉綠素與酚類化合物造成影響，進而改變兩者之螢光反應。因此本研究建立一包含UV-A為激發光源，且能同步擷取四通道影像之多光譜螢光影像系統，其中特徵波長以遙測光譜儀GER 2600搭配UV-A為激發光源之實驗分析所確定。本研究成功選擇出530 nm、680 nm與740 nm為蘋果碰撞造成褐變之相關特徵波長，並以累加連續時間拍得之螢光影像，達到類似增加曝光時間又不喪失時間軸解析度的方法。本研究將上述三個波長下擷取之螢光影像隨時間轉換成螢光強度曲線，以主成份與PLS-DA進行定性分析，主成份分析結果中，蘋果組織褐變之螢光差異性於波長530 nm最大，PC-1變異量即達99.79 %，累計至PC-3，變異量更高達99.95 %；而以PLS-DA定性分析預測蘋果組織褐變與否，於波長530 nm結果最佳，累計前四個LV的變異量可達97.59 %，足以100 %正確預測蘋果組織褐變。

The quality inspection of fruits is a major issue on agricultural development; it can not only protect consumer’s rights, but also improve farmers' income. There are chemical changes inside the fruit when internal browning occurs, flesh color will be changed into dark brown after browning. There are many situations may cause fruit browning, for examples, stored in improper controlled atmosphere, impact during transport, etc. Above mentioned browning are caused by polyphenoloxidase, which is contained in fruit and will induce the oxidation of phenolic compounds, then is polymerized into insoluble dark brown polymer (melanins). Browning will reduce the economic value of fruits, using non-destructive sorting of fruits can ensure the quality of fruits. Chlorophyll fluorescence measurement is a non-destructive method. Recently, there are two-dimensional multispectral fluorescence imaging system used for internal quality inspection of bio-materials. Fruit browning which is caused by impact will change the fluorescence responses. A multispectral fluorescence imaging system with UV-A light as an excitation light source was developed in this study, and was able to capture four images simultaneously. The multispectral bands (530, 680, 740 nm) used in the system can discriminate apples browning. Accumulated consequent fluorescent images to increase fluorescence intensity can overcome the problem of inadequate sensitivity of camera. This study transformed fluorescent images into fluorescent intensity curve, and principal component analysis (PCA) and PLS-DA (Partial Least Squares - Discriminant Analysis) were used for qualitative analyses. Principal component analysis showed that fluorescence at 530 nm had maximum difference of tissue browning, the variance of PC-1 amounted to 99.79 %, and to 99.95 % (accumulated PC-1 to PC-3). PLS-DA also showed best results when fluorescence was measured at 530 nm, and was able to predict browning of apple tissues.