Application of Multivariate Analysis on Digital Images of Cannabis Sativa L Extracts

Abstract

Cannabis sativa L is one of the most used drugs in the world. Information about the plant’s age and storage can help forensic scientists to identify and to track samples. The ratio between the cannabinoids tetrahydrocannabinol (THC) and cannabinol (CBN) has been related to the degradation of cannabis with time. Thus, this study aimed to test Multivariate Image Analysis (MIA) to evaluate cannabis extracts concerning its colors. Initially, 52 samples of Cannabis sativa L. extracts were analyzed by Gas Chromatography coupled to Flame Ionization Detector (GC/FID) to quantify THC and CBN. Afterwards, the extract samples were photographed and analyzed by two different multivariate analysis tools: ChemoStat®, a free chemometrics software, and PhotoMetrix PRO®, an app for mobile devices. Using exploratory analysis of principal component analysis (PCA) and hierarchical cluster analysis (HCA). It was observed that the more intense the color for an extract, the higher concentration of THC and CBN it has, while the lighter color extracts correspond to samples with no THC. The results suggest to propose a simple method for previous clustering of samples that may precede chromatographic analyzes, assist in chemical profile studies or simply aggregate samples of similar profiles for analyzed together.