Detection of Adulteration in cinnamon powder using hyperspectral imaging

Document Type : Research Paper

Authors

• mohamadhossein nargesi ¹
• Jafar Amiri Parian ²
• hossein bagherpour ³
• Kamran Kheiralipour ⁴

¹ Biosystem Mechanical Engineering Department, Faculty of Agriculture, Ilam University, Ilam, Iran.

² Department of Biosystems Engineering, Faculty of Agriculture, Bu-Ali Sina University, Hamedan, Iran

³ Department of Biosystem Engineering, Faculty of Agriculture, Bo-Ali Sina University, Hamadan, Iran.

⁴ Department of Biosystem Engineering, Faculty of Agriculture, Ilam University, Ilam, Iran.

Abstract

Keywords

• Cinnamon
• Adulteration
• Hyperspectral imaging
• Image processing
• Machine learning

Main Subjects

• pardazesh tasvir

Determining the purity of cinnamon powder using hyperspectral imaging

EXTENDED ABSTRACT

Introduction

Cinnamon, an evergreen plant, is one of the most important food seasonings and spices that has been used for thousands of years as medicinal plants in the treatment of diseases. It has antioxidant and medical properties that regulates immune system and is an anti-inflammatory against various diseases. Hyperspectral imaging as a new technique to assess the quality and purity of agricultural and food products. The purpose of this study is to determine the purity of cinnamon powder using hyperspectral image processing technique.

 Materials and Methods

The present research was done in the image processing laboratory of Ilam University, Ilam, Iran. Five levels of impurities including 0, 5, 15, 30 and 50% were considered to be determined by the system. Impurity materials were wheat flour, chickpea flour, and sea foam powder. For each impurity level, three samples were provided and kept in bags. The image was acquired using a line scanning hyperspectral imager. Six hyperspectral images were acquired from each sample so that 18 hyperspectral images were acquired from each impurity level so that 270 hyperspectral images were obtained for each impurity material. MATLAB software was used to analyze hyperspectral images. Image processing step included wavelength selection, feature extraction, and feature selection. The efficient features were classified using the support vector machine method.

 Results and Discussion

The confusion matrixes of the classifier model based on support vector machine method with one-against-one and one-for-all strategies were obtained to calculate the correct classification rates of the models. The correct classification rates of the classifier with one-to-one strategy for detecting chickpea flour, wheat flour, and sea foam powder impurity in cinnamon powder were 95.55, 96.66, and 85.56%, respectively. The correct classification rates of the model with one-against-all strategy were 78.88, 94.44, and 77.77%, respectively.

Conclusion

The results of the present study showed the high ability of hyperspectral imaging technology combined with support vector machine classifier method with one-to-one strategy in detecting wheat flour, chickpea flour, and sea foam impurities in cinnamon powder. The proposed methodology in the present research has different advantages over laboratory-based methods, including non-destructiveness, high speed, and low cost. It is suggested to use other methods to classify hyperspectral images in order to detect impurity in cinnamon. The proposed method in the present research can be used in the future to detect other types of fraud in cinnamon.