Detection of Apple Defects Using an Electronic Nose and zNose
電子鼻技術用于蘋果質量缺陷檢測
Changying Li, PhD Student, Pennsylvania State University, 249 Agricultural Engineering
Building, University Park, PA, 16802. cul140@psu.edu
Paul Heinemann, Professor, Pennsylvania State University, 249 Agricultural Engineering
Building, University Park, PA, 16802
Joseph Irudayaraj, Associate Professor, Purdue University, 225 S. University St., West
Lafayette, IN 47907
Devin Peterson, Assistant Professor, Pennsylvania State University, 215 Borland Laboratory,
University Park, PA, 16802
Written for presentation at the
2005 ASAE Annual International Meeting
Sponsored by ASAE
Tampa Convention Center
Tampa, Florida
17 - 20 July 2005
Abstract. Apple defects and spoilage not only reduce commodity economic value, but cause food safety concerns as well. It is essential for fruit quality assurance and safety to rapidly detect fruit physical damage and spoilage. This article presents the application of an electronic nose (Cyranose 320) and zNose to the development of a nondestructive, rapid and cost effective system for the detection of defects of apples. The key compounds associated with apple aroma were identified and the “smellprints” of these key compounds were established by the electronic nose and zNose. Healthy and damaged apples were kept in 2L glass jars for 6 hours for preconcentration before measuring. Principal Component Analysis (PCA) models were developed based on the Enose and zNose data. Maholanobis distance was applied for discriminant analysis. Experiments showed that the Enose and zNose are both capable of detecting the volatile differences between healthy apples and damaged apples. After five days deterioration, the correct classification rate for the Enose was 83.3%, and for the zNose was 100%. After seven days, the correct classification rate was 100% for both instruments. For the next stage, a non-linear model and sensor fusion technique will be developed.
蘋果的缺陷和變質不僅降低了商品的經濟價值，也引起了食品安全的擔憂。快速檢測水果的物理損傷和變質是保證水果質量和安全的關鍵。本文介紹了電子鼻(Cyranose320)和zNose在開發無損、快速、經濟有效的蘋果缺陷檢測系統中的應用。通過電子鼻和zNose，鑒定了與蘋果香氣有關的關鍵化合物，并建立了這些關鍵化合物的“氣味圖譜”。健康和受損的蘋果在2升玻璃罐中保存6小時，在測量前進行預濃縮。基于ENOSE和ZNOSE數據建立了主成分分析(PCA)模型。采用Maholanobis距離進行判別分析。實驗表明，該酶和氧化鋅均能檢測健康蘋果和受損蘋果的揮發性差異。經過5天的變質處理后，eNose的正確分類率為83.3%，Znose的正確分類率為100%。7天后，兩種儀器的正確分類率為100%。下一階段將開發非線性模型和傳感器融合技術。
