Determining the purity of black pepper powder by hyperspectral imaging method

Introduction: Plants and spices are the source of many biologically active substances that can enhance the taste, color, and aroma of food, as well as influence the body's digestion and metabolic processes. A spice is a dried seed, fruit, root, bark, or vegetable substance primarily used for seasoning, coloring, or preserving food. Sometimes, spices are also used to mask other flavors. On the one hand, many spices have antimicrobial, anti-diabetic, anti-inflammatory, and anti-hypertensive properties. Black pepper, in particular, has been used as a pain reliever in traditional medicine for centuries. This plant has been cultivated since ancient times both as a spice and as a medicinal herb, and it has also been a significant commercial product. Emerging scientific techniques, such as hyperspectral imaging, are now used to evaluate the quality and purity of agricultural and food products. The purpose of this study is to determine the purity of black pepper powder using hyperspectral image processing techniques.Materials and Methods: The line scan camera from the university's image processing workshop was used to conduct this research. Adulterants such as wheat flour, peas, and sea foam were mixed with black pepper powder at impurity levels of 0%, 5%, 15%, 30%, and 50%. Three samples were prepared for each level of impurity and stored in zip bags. Six images were recorded from each sample. A total of 270 hyperspectral images were recorded. MATLAB software was used to analyze these images. The samples underwent pre-processing, which included the selection of length, features, and characteristics. Efficient features were then classified using the support vector machine method.Results and discussion: The confusion matrices of the support vector machine classifier model were calculated using one-for-one and one-for-all strategies to determine the correct classification rate for black pepper fraud detection. The accuracy of the support vector machine classification model with the one-against-one strategy in detecting adulteration with wheat flour, sea foam, and chickpea flour in black pepper was 98.88%, 98.88%, and 95.55%, respectively. Using the one-against-all strategy, the accuracy was 100%, 91.11%, and 93.33%, respectively.Conclusions: In the present study, the classification of different levels of adulteration in black pepper was performed using the hyperspectral image processing method and support vector machine. Due to the varying levels of adulteration, two strategies were employed: one-against-one and one-against-all, with the one-against-one strategy yielding better performance. Besides, this research method offers several advantages over traditional laboratory methods, including non-destructiveness, high speed, and low cost. It is suggested to explore other classification methods for hyperspectral images to further improve the detection of impurities in black pepper.