Skip to main content
ARS Home » Pacific West Area » Corvallis, Oregon » Horticultural Crops Production and Genetic Improvement Research Unit » Research » Publications at this Location » Publication #427505
Research Project: Resilient Production Strategies for Improved Small Fruit Quality

Location: Horticultural Crops Production and Genetic Improvement Research Unit

Title: A systematic color correction pipeline for controlled-environment imaging

Author
item WAKHOLI, COLLINS - Oak Ridge Institute For Science And Education (ORISE)
item Hardigan, Michael
item Lee, Jungmin
item LUKAS, SCOTT - Oregon State University
item Feldman, Maximilian
item Altendorf, Kayla
item Neyhart, Jeffrey
item Rippner, Devin

Submitted to: The Plant Phenome Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/4/2026
Publication Date: 3/11/2026
Citation: Wakholi, C., Hardigan, M.A., Lee, J., Lukas, S.B., Feldman, M.J., Altendorf, K.R., Neyhart, J.L., Rippner, D.A. 2026. A systematic color correction pipeline for controlled-environment imaging. The Plant Phenome Journal. 9(1). Article e70067. https://doi.org/10.1002/ppj2.70067.
DOI: https://doi.org/10.1002/ppj2.70067

Interpretive Summary: Scientific color analysis in digital images is challenging because perceived color in images is dependent on lighting, background color, subject surface properties, and camera properties. A color correction pipeline was developed to enable accurate color correction in digital images. This pipeline was tested for robustness using different lights, light positions, backgrounds, and cameras using color cards in different positions. The results show improved color accuracy with the use of the color correction pipeline under all conditions.

Technical Abstract: A comprehensive, stepwise color correction pipeline designed especially for controlled imaging environment applications is presented. The proposed pipeline integrated Flat Field Correction (FFC), Gamma Correction (GC), and White Balancing (WB) with machine learning regression—including linear, partial least squares (PLS), and neural network (NN) approaches—for reliable color correction in digital images. Experimental evaluations revealed that high-quality illumination and their positioning are essential for achieving consistent color reproduction. Although the commonly used 45° illuminant position produced favorable uncorrected images, top-mounted area illumination combined with FFC yielded superior performance. The stepwise image processing reduced perceptual color difference in the corrected images. Using a NN-based fitting method with a 2-degree polynomial expansion outperformed all other fitting approaches that delivered the lowest color difference, and demonstrated robust performance across variable imaging conditions. Imaging environment variables such as object background color and side wall finishes were also found to impact color fidelity, emphasizing the need for careful selection of these setup materials. This study underscores the potential of an integrated, open-source color correction pipeline to improve color reproduction and measurement by digital imaging. It also bridges the gap between sophisticated color correction methods and practical applications for future studies.