HPLC and Image Analysis of Carrot Pigments

Analyzing the color of food using digital images is a simple, non-damaging way to estimate the levels of certain natural pigments responsible for red, orange, and yellow colors in plants. While existing research has produced reasonably accurate models for predicting these pigment levels from color data, they tend to be basic. What's missing are more robust, standardized approaches that account for the fact that color measurements are often closely linked to one another, that lighting conditions can vary, and that the foods being analyzed can differ widely in their genetic makeup.

Researchers at Brightlands Future Farming Institute at Maastricht University (Venlo, Netherlands) developed an improved image analysis method to predict pigment levels in 16 different carrot varieties spanning a range of colors. Each sample was photographed under six different lighting conditions using a digital camera, and the images were color-corrected before analysis. Total pigment contents and individual carotenoid contents were analyzed chemically via spectrophotometry and high-performance liquid chromatography (HPLC), respectively. A specialized statistical method was then used to explore the relationship between the color data and pigment levels, while accounting for the fact that the various color measurements tend to be closely related. A paper based on this work was published in Frontiers in Plant Science.¹

Why Do Pigments Matter Regarding Fruits and Vegetables?

Carotenoids and anthocyanins are two major groups of natural pigments that give fruits and vegetables their vibrant colors. Carotenoids are responsible for yellow, orange, and some red shades, while anthocyanins produce other reds and purple-blue tones.^2,3. Both pigment groups are powerful antioxidants that work in multiple ways to support health, making them valuable components of a healthy diet. Beyond their health benefits, vibrant colors in food also stimulate appetite and aid digestion, and research has shown that consumers often judge the quality of fresh produce based on its color.^4,5

What Did the Study Find, and What Are the Next Steps?

The researchers report that their models performed well at predicting overall levels of both orange-yellow and red-purple pigments across all lighting conditions, correctly accounting for around 77–81% of the variation seen. These models are considered reliable even when lighting varies and when working with a genetically diverse range of samples. However, predicting the levels of individual pigments within each carrot variety proved more difficult, largely because the carrots showed a great deal of genetic variation even within the same variety. Despite this, the findings are still valuable because the data can be used to build larger databases that could support artificial intelligence applications, and to help plant breeders develop carrot varieties with higher levels of these beneficial antioxidants.¹

The team believes that the scientific community needs to agree on a consistent approach to taking and processing food images, as this plays a major role in how accurately pigment levels can be predicted and helps avoid misleading results. This includes using color reference cards and testing how different lighting affects the outcome. It is also recommended to directly measure the light hitting the sample, use larger and more varied sample groups, and take more color readings per sample to improve the models further. Better understanding of how genetics and environment influence pigment levels could also be explored.¹

The researchers pointed out that a notable aspect of their study was that it was the first to set aside environmental variation in carrots when looking at their pigment combinations. They suggest that future work take a broader approach, looking at multiple plant species and how their surroundings affect pigment levels. Expanding these models to other foods and building larger databases could eventually train artificial intelligence to automatically detect pigment levels, which could form the foundation of future food and farming technologies. For example, this method could one day be built into a phone app that uses artificial intelligence (AI) to estimate the antioxidant content of fresh produce just from a photo. This could help retailers and hospitals make better, more informed decisions about food selection and reduce unnecessary waste.¹

Read More on Similar Topics
Pathway-Based Metabolomics Reveals Biosynthetic Routes of Flavor and Pigment Compounds in Pomegranate

References

1. Verme, A. C.; Melchior, M.; van den Berg, J. H. et al. Image Analysis Optimisation for Carotenoid and Anthocyanin Content Prediction in Carrots: Addressing Colour Parameter Multicollinearity and Genotypic Diversity. Front Plant Sci. 2026, 17, 1713048. DOI: 10.3389/fpls.2026.1713048
2. Amorim-Carrilho, K. T.; Cepeda, A.; Fente, C. et al. Review of Methods for Analysis of Carotenoids. TrAC Trends Anal. Chem. 2014, 56, 49–73. DOI: 10.1016/j.trac.2013.12.011
3. HolmeI, B.; Dionisio, G.; Brinch-Pedersen, H. A Roadmap to Modulated Anthocyanin Compositions in Carrots. Plants 2021, 10, 472. DOI: 10.3390/plants10030472
4. Munsch, M. H.; Simard, R. E.; Girard, J.-M. Relationships in Colour and Carotene Content of Carrot Juices. Can. Inst. Food Sci. Technol. J. 1983, 16, 173–178. DOI: 10.1016/S0315-5463(83)72203-0
5. Pérez, M. B.; Da Peña Hamparisomian, M. J.; Gonzalez, R. E. et al. Physicochemical Properties, Degradation Kinetics, and Antioxidant Capacity of Aqueous Anthocyanin-Based Extracts from Purple Carrots Compared to Synthetic and Natural Food Colorants. Food Chem. 2022, 387. DOI: 10.1016/j.foodchem.2022.132893