Multivariate Classification of Prunus dulcis Varieties using Leaves of Nursery Plants and Near-Infrared Spectroscopy

Abstract

The emergence of new almond tree (Prunus dulcis) varieties with agricultural interest is forcing the nursery plant industry to establish quality systems to keep varietal purity in the production stage. The aim of this study is to assess the capability of near-infrared spectroscopy (NIRS) to classify different Prunus dulcis varieties as an alternative to more expensive methods. Fresh and dried-powdered leaves of six different varieties of almond trees of commercial interest (Avijor, Guara, Isabelona, Marta, Pentacebas and Soleta) were used. The most important variables to discriminate between these varieties were studied through of three scientifically accepted indicators (Variable importance in projection¸ selectivity ratio and vector of the regression coefficients). The results showed that the 7000 to 4000 cm−1 range contains the most useful variables, which allowed to decrease the complexity of the data set. Concerning to the classification models, a high percentage of correct classifications (90–100%) was obtained, where dried-powdered leaves showed better results than fresh leaves. However, the classification rate of both kinds of leaves evidences the capacity of the near-infrared spectroscopy to discriminate Prunus dulcis varieties. We demonstrate with these results the capability of the NIRS technology as a quality control tool in nursery plant industry.

A Review of Machine Learning for Near-Infrared Spectroscopy

“…The exhaustive search required by wrapper-based methods to prevent overfitting also makes them computationally expensive. Wrapper-based algorithms used in NIR include particle swarm optimization (PSO) and binary particle swarm optimization (BPSO) [28], genetic algorithms (GA) [29][30][31][32][33], variable combination population analysis (VCPA) [32,34], the variable iterative space shrinkage approach (VISSA) [35], bootstrapping soft shrinkage (BOSS) [36], iteratively retaining informative variables (IRIV) [32], competitive adaptive reweighted sampling (CARS) [37][38][39], the successive projection algorithm (SPA) [40], uninformative variable elimination (UVE) [41], Monte Carlo uninformative variable elimination (MCUVE) [35], partial least squares feature selection approaches [42], the randomization test (RT) [43], variable importance in the projection (VIP) [44], and the jackknife procedure.…”

Geographic Origin Discrimination of Millet Using Vis-NIR Spectroscopy Combined with Machine Learning Techniques

“…NIRS has already been established in the 1960s for cereal analyses [18]. Geographical origins of many kinds of cereal and other agricultural products have been determined using NIR in recent years, such as maize [19], walnuts [20], durum wheat [21], rice [22][23][24], turmeric [25], kudzu powder [26], Prunus Dulcis [27], Trichosanthis Fructus [28], Chinese mitten crab [29], edible oils [30], Wolfiporia cocos [31], Argentinean lemon juices [32], honey [33,34], etc.…”

Varietal quality control in the nursery plant industry using computer vision and deep learning techniques

“…Other studies have focused on the implementation of near-infrared (NIR) spectroscopy as a tool for varietal control. 17,18 In the literature, there are studies related to plant identification using CV. [19][20][21][22] In our case, the challenge is to discriminate between two almond trees (Prunus dulcis) varieties, Soleta and Pentacebas, which are morphologically very similar, genetically close, and almost indistinguishable for the human eye.…”

“…Despite these molecular techniques have an excellent accuracy, they are very expensive for routine analysis of a large number of samples. Other studies have focused on the implementation of near‐infrared (NIR) spectroscopy as a tool for varietal control 17,18 …”