Spectral signatures: A way to identify species and conditions of vegetables

Abstract

This work shows the results obtained by the application of suitable techniques for the acquisition and processing of spectral absorbance data of several vegetable species, allowing its identification –assignment of each spectral signature to one specific plant- which in turn allows the control of origin of products (foodstuffs or not) and their characteristics. Spectral measurements of absorbance were performed on samples of two sugarcane varieties and four citrus types (orange, lemon, tangerine and grapefruit) taken periodically from a controlled crop, using a spectrophotometer FOSS-NIR 6500 in the range of 400-2500 nm by 2 nm. The measured samples were about 240 in the case of sugarcane and 160 in the case of citrus. A Principal Component Analysis was applied to the data by means of STATA 9 software and the results were interpreted in that PCA context. This procedure allowed us not only to clearly identify which variety of sugarcane corresponds to each spectral absorbance function, but also to determine which wavelength or wavebands have significant relevance for that identification. Similarly, this technique allows us to identify and classify the spectral functions coming from different types of citrus. The main conclusion is that the proposed technique is capable to precisely identifying the species each sample comes from; besides, this technique would allow us to determine the nutritional or health condition of them at the moment of the analysis, as could be seen from the obtained results.