Pathange, Lakshmi Prasad2014-03-142014-03-142003-02-06etd-02212003-092504http://hdl.handle.net/10919/41235The apple growers and packaging houses are interested in methods that can evaluate the quality of apples non-destructively. Harvested fruits are a mixture of immature, mature, and over mature fruits, thereby posing a great problem in deciding their end use and storage time. It is expected that the technique developed from the present project could be effectively used to classify the harvested fruit into immature, mature and over mature apples, rapidly and non-destructively. It would also help the growers to predict the optimum dates to harvest the fruits. York and Gala were the varieties of apples that were used in this study and were obtained from Virginia Tech College of Agriculture and Life Sciences Kentland Farm. Apples were harvested at different times resulting in different maturity groups (immature, mature and ripe). Gala apples were harvested on three dates with an interval of 10 days, while York apples were harvested on four dates with an interval of 14 days. They were stored at 0oC until sampled. For each harvest date, the experiments were conducted in two sets (10 each) on two consecutive days. First the ethylene levels were measured, followed by gas chromatograph and electronic nose. Then the maturity indices were measured. Three maturity indices, starch index, firmness and soluble solids were used as the three variables for the statistical analysis to identify and categorize the fruits into three maturity categories referred as immature, mature and over mature fruits. Apples were also categorized into three maturity groups based on the emanation levels of the aroma compounds evolved from the fruits. Then electronic nose sensor responses were categorized into the above maturity categories, and their effectiveness was determined using a statistical procedure called Discriminant Analysis (DA). From the DA cross validation results the correct classification percentage for Gala and York apples into maturity groups was 95%. The Electronic nose sensor's effectiveness to categorize the same observations based on sensor responses in to the above classified maturity categories was 83% correct in case Gala apples and 69% for York apples. The EN sensors response data were analyzed by the EN system software and the correct classification percentage for Gala was 83% and for York was 81%. Aroma-based categorization for Gala apples was 100% correct, while the electronic nose for the same analysis was 80%. Based on the three physical parameters, an objective evaluation of maturity could be accomplished. Principal Component Analysis, Canonical Discriminant Analysis and DA results demonstrated that the electronic nose could be used to classify apples into three identified maturity-based groups. The EN sensors (Gala apples), could also classify the apples into aroma-based categories. Thus, it can be concluded that the EN system holds promise as non-destructive evaluation technique to determine the maturity of an apple.In CopyrightYorkStatistical AnalysisGalaApplesElectronic NoseMaturity IndicesThesishttp://scholar.lib.vt.edu/theses/available/etd-02212003-092504/