Introduction: Saffron (Crocus sativus L.) is one of the most valuable spices worldwide because of its distinctive color, aroma, and flavor. Its high economic value and limited production have made it particularly susceptible to adulteration. The addition of synthetic dyes is among the common fraudulent practices used to intensify the apparent color of saffron and increase its commercial value. Therefore, the development of rapid, reliable, and accessible methods for detecting artificial colorants is important for saffron quality control and food authenticity assessment. Conventional analytical methods such as chromatography and spectrophotometry can provide valuable information for identifying adulterants; however, some of these techniques require specialized equipment, considerable analysis time, or complex sample preparation. Digital image processing combined with chemometric techniques has emerged as a promising alternative for rapid and non-destructive evaluation of food quality. Principal component analysis (PCA), in particular, can reduce the dimensionality of multivariate color data and reveal differences among samples. Therefore, the present study aims to investigate the applicability of thin-layer chromatography (TLC), spectrophotometry, digital image processing, and PCA for the detection and discrimination of saffron samples adulterated with synthetic dyes.

Materials and Methods: Four synthetic food dyes, including tartrazine, Erythrosine, quinoline yellow, and sunset yellow, were investigated. Stock solutions of the dyes were prepared at a concentration of 1 g/L, followed by preparation of working solutions at 0.001 and 0.0005 g/L. Saffron samples were artificially adulterated using individual dyes, binary combinations, and a mixture containing all four dyes, while untreated saffron was used as the control. For this purpose, 5 g portions of saffron were treated with the corresponding dye solutions, dried, homogenized, and stored at 4°C until analysis. For extraction, 0.5 g of each sample was mixed with 25 mL of distilled water at 60°C, shaken for 1 min, and centrifuged at 4000 rpm for 10 min. The resulting-colored aqueous phase was collected for further analysis. TLC was employed for preliminary identification of synthetic dyes based on their retention factor (RF) values. Spectrophotometric measurements were performed at the maximum absorption wavelength (λmax) of each dye. For digital image analysis, saffron samples were photographed under controlled illumination using a fixed camera position and identical imaging conditions. The obtained images were analyzed using ImageJ software, and color parameters were extracted from the CIELAB, HSV, and RGB color spaces. The resulting variables included L*, a*, b*, B1, S, G, and B2. Variables H and R were excluded from subsequent analysis because of their very low contribution to the explained variance. PCA was performed using Minitab 16 to investigate the multivariate relationships among color parameters and to visualize the discrimination of control and adulterated samples.

Results and Discussion: Comparison of the Rf values obtained by thin-layer chromatography (TLC) with those of the reference dyes provided useful qualitative information for identifying artificial colorants in saffron samples. Spectrophotometric analysis also revealed distinct absorption characteristics for the investigated dyes. In particular, the sample containing the mixture of all four dyes exhibited a considerably different spectral response compared with the other treatments, whereas some individual and binary adulterations showed relatively similar responses to the control sample. Digital image analysis provided additional information regarding changes in the visual color characteristics of adulterated saffron. PCA demonstrated that the extracted color parameters contained sufficient information to distinguish several adulterated samples from the control sample. At a dye concentration of 0.001 g/L, the first and second principal components had eigenvalues of 5.2665 and 0.9637 and explained 75.2% and 13.8% of the total variance, respectively, accounting for approximately 89.0% of the cumulative variance. At 0.0005 g/L, PC1 and PC2 explained 91.6% and 5.1% of the total variance, respectively, with a cumulative variance of approximately 96.7%. These results indicate that most of the variability in the image-derived color data could be represented by only two principal components. The score plots generally showed separation between the authentic saffron and several adulterated groups, although some overlap was observed among particular dye treatments and combinations. These findings indicate that PCA based on color parameters can be useful for exploratory discrimination and rapid screening, but it may not be sufficient for the unequivocal identification of every adulterant. Similar studies have demonstrated the potential of chemometric methods, including PCA and supervised classification techniques, for detecting synthetic colorants and other forms of saffron adulteration.

Conclusion: The proposed method is a simple and potentially low-cost approach for the preliminary screening of saffron adulterated with synthetic dyes. The high proportion of explained variance by the first two principal components confirms the effectiveness of the extracted color parameters for describing differences among samples. Nevertheless, because some adulterated groups exhibited partial overlap, image-based PCA should preferably be considered a screening tool rather than a definitive identification method. Combining digital image analysis with complementary techniques such as TLC and spectrophotometry can improve the reliability of saffron authenticity assessment and provide a practical strategy for quality control.