Feasibility of Visible/Near Infrared (Vis/NIR) Spectrophotometry capability in classification of lemon samples during storage period by PCA, LDA and SVM identification methods

Today, the increasing process of food waste and agricultural products is one of the serious challenges in the most countries, especially in developing countries, so one of the serious policies of governments in the food security is to reduce the waste and maintain the quality of agricultural products. So far, several methods have been used to measure the quality of agricultural products, only some of which are technically and industrially justified. Vis / NIR Spectrophotometry method is one of the methods that has been considered and used in evaluating the qualitative characteristics of agricultural products due to its high speed and accuracy. In this regard, in the present study, visible/near infrared Spectrophotometry was used to measure the qualitative changes and classification of K-Lime samples of lemon during the storage period (10, 20 and 30 days). In order to analyze the qualitative characteristics and classify the data extracted from NIR, the pattern recognition methods including principal component analysis (PCA), linear Discriminant analysis (LDA) and support vector machine (SVM) were used. The results showed that Visible/Near Infrared (Vis/NIR) Spectrophotometry was able to differentiate its lemon samples based on storage time. Although PCA, LDA and SVM methods were able to classify lemon samples with good accuracy according to qualitative characteristics, but LDA and SVM methods with 100% accuracy had better accuracy and fit. Also, according to the results, the quadratic function has been determined and introduced as the best function for constructing classification models by LDA and SVM methods.