Loading...

Determination of Saffron Adulteration Thorough the Package Using Hyperspectral Imaging and Chemometric Techniques

Ostovar, Mona | 2022

232 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 55118 (03)
  4. University: Sharif University of Technology
  5. Department: Chemistry
  6. Advisor(s): Parastar Shahri, Hadi
  7. Abstract:
  8. These days, food authenticity has become a major challenge because food health directly affects human health. The importance of authenticity is highlighted when we are faced with foods with higher nutritional and economic value. Saffron is an important example of spices because in addition to having food coloring and flavoring properties, it also has numerous health benefits, but it has limited production and high price. Therefore, with the options available and cheaper to be replaced. Thus, among the various spices, cheating in saffron has the fourth place. Hyperspectral Imaging (HSI) has been developed for food safety industrial applications. This technique combines spectroscopy and imaging techniques, enabling the acquisition of hyperspectral images. These images are composed of reflectance pixelmaps containing hundreds of channels, one per wavelength. Hyperspectral imaging can be considered as a promising solution for detecting food contamination. Since Iran is the largest producer of saffron in the world and the scope of fraud in this food is wide, in the present project, the authenticity of saffron in the form of packaging in the field of common frauds, based on hyperspectral imaging method has been investigated. On the other hand, the analysis of the obtained fingerprints, considering the complexity of saffron texture and the size and complexity of the responses of the hyperspectral imaging technique, requires the use of chemometric methods. For this purpose, In this study, hyperspectral imaging along with chemometrics methods has been used as a data analysis tool to authenticate and detect 3 common adulterants in Iranian saffron samples (safflower, calendula, style). The importance of this project is to check the authenticity of saffron commercially. (packaged and without the need for pulverization). In this study, four data sets have been examined, In each dataset, the multivariate curve resolution-alternative least squares algorithm was first used to achieve pure spectral and concentration profiles. Authentication was performed based on spectral profiles and classification was performed based on concentration profiles. Dataset I, Includes 38 authentic samples of Iranian saffron from different regions of Khorasan Razavi. Using the results of spatial distribution map analysis performed by the Principal Component Analysis (PCA) algorithm, six authentic samples of saffron were selected as pooled sample for further analysis. Dataset IIa, IIb, IIIc were for binary mixture of pooled saffron sample and safflower, calendula and style adulterants respectively in five level of 5, 10, 15, 25, 35% (w/w). For classification purpose Partial Least Squares-Discriminant Analysis (PLS-DA) algorithm wase used on spatial profiles, and the accuracy results for all dataset IIa, IIb, IIc were 100 percent. Dataset III, Includes mixture of all adulterants and pooled saffron sample. Accuracy for classification results using PLS-DA algorithm on spatial profiles were 100 %. Dataset IV, includes a binary set consisting of saffron and safflower that has been studied with three different types of packaging. Accuracy of the classification results using Partial Least Squares-Discriminant Analysis (PLS-DA) algorithm on spatial profiles: IVa) Compressed plastic cover, accuracy=93.3; IVb) Thick plastic cover, accuracy= 97.6; IVc) Thin plastic cover, accuracy=100.The results showed that the developed model can be used as a standard method to check the authenticity of saffron in packaging without the need for pulverization
  9. Keywords:
  10. Saffron ; Hyperspectral Images ; Chemometrics Method ; Multivariate Curve Resolution ; Skulduggery ; Principal Component Analysis (PCA)

 Digital Object List

 Bookmark

No TOC