The system, tested on eight lettuce species, successfully visualized the spatial distribution of pigments from individual leaves to the entire canopy, providing a powerful new tool for precision crop management and physiological monitoring.
Plant pigments are important indicators of photosynthesis, nutrition, and overall plant health. Lettuce, a widely consumed leafy vegetable, contains high levels of photosynthetic pigments that are essential for growth, flavor, antioxidant capacity, and nutritional value. Chlorophyll promotes light absorption and carbohydrate synthesis, and carotenoids protect tissues from oxidative damage and serve as precursors for vitamin A in humans. Traditional dye quantification relies on solvent extraction or chromatography, which, while accurate, is destructive and not suitable for continuous field evaluation. In recent years, hyperspectral sensing and machine learning have enabled nondestructive nutrient detection, but most approaches require complex feature preprocessing, lack cross-scale evaluation, and are difficult to generalize across varieties. These challenges highlight the need for rapid, automated, and scalable methods for assessing pigments in living plants.
a study (DOI: 10.1016/j.plaphe.2025.100104) Published in plant phenomics September 12, 2025 Li-Ping Chen's team at the Beijing Academy of Agricultural and Forestry Sciences uses hyperspectral imaging combined with deep learning to establish a highly accurate and non-destructive method to estimate and visualize the pigment content of lettuce, enabling efficient physiological assessment and supporting precision agriculture.
In this study, the authors first performed a quantitative analysis of pigment content in the leaves of eight lettuce varieties, combined with hyperspectral reflectance measurements and a series of modeling approaches, to investigate how pigments vary within and between cultivars and how accurately they can be predicted from spectral data. They characterized the distributions of Chl a, Chl b, Car, and TPC, analyzed the vis-NIR reflectance curves (366–976 nm) and red edge features, and preprocessed the hyperspectral data using MA, SNV, and FD1 to enhance the spectral peaks and valleys. Machine learning models were systematically built based on different combinations of preprocessing, feature selection (including CARS and SPA), dataset partitioning (random, KS, SPXY), and regression algorithms (PLSR, RF, SVR, ELM). Then, an end-to-end deep learning model, LPCNet (CNN + BiLSTM + MHSA), was trained and evaluated. Finally, we extended the leaf-level inversion model to canopy reflectance to visualize pigment maps and used canopy pigment statistics to compare distribution patterns among lettuce species. The results showed different statistical distributions for each dye. Chl a was bimodal (0.68–0.99 mg/g), Chl b was unimodal (0.29–0.45 mg/g), Car was right-biased (0.14–0.23 mg/g), and TPC was nearly normal (1.19–1.55 mg/g), reflecting diverse metabolic strategies. Spectral analysis confirmed characteristic absorption valleys and pronounced red edges at 430–470 nm and 670–690 nm, and pretreatment improved signal stability. While the best machine learning combination achieved high R² values (up to ~0.91), LPCNet further improved accuracy and robustness, reaching R² up to 0.94 and lower MAE on the prediction set. Pseudo-color canopy maps revealed spatial gradients of Chl a, Chl b, Car, and TPC and consistent patterns among lettuce types as well as clear differences between lettuce types, linking pigment distribution to genetic background, leaf morphology, and light adaptation strategies, and provided a solid foundation for breeding and cultivation optimization.
This study establishes a fast, non-invasive pigment estimation pipeline suitable for high-throughput phenotyping, smart greenhouses, and digital farm monitoring. Real-time pigment maps allow growers to assess photosynthetic efficiency, diagnose nutritional stress early, optimize fertilizer strategies, and support variety selection based on pigment traits. For research, this system provides a new lens to investigate pigment metabolism, light response mechanisms, and biochemical changes in plants under environmental changes.
###
References
Toi
10.1016/j.plaphe.2025.100104
Original source URL
https://doi.org/10.1016/j.plaphe.2025.100104
Funding information
This research was partially supported by the Beijing Rural Revitalization Agricultural Science and Technology Project (NY2401040025), the National Key Research and Development Program (2022YFD2002300), and the Construction of Collaborative Innovation Center of the Beijing Academy of Agricultural and Forestry Sciences (KJCX20240406).
About plant phenomics
plant phenomics is dedicated to publishing new research that advances all aspects of plant phenotyping from the cellular to the plant population level using innovative combinations of sensor systems and data analysis. Plant Phenomics also aims to connect phenomics to other scientific areas such as genomics, genetics, physiology, molecular biology, bioinformatics, statistics, mathematics, and computer science. plant phenomics Therefore, there is a need to contribute to the advancement of plant science and agriculture/forestry/horticulture by addressing important scientific challenges in the field of plant phenomics.
