Title : An integrated RGB and spectral phenotyping framework for rice seedling salt-stress evaluation and genetic mapping
Abstract:
Rice seedling salt tolerance is commonly assessed using visual scores or survival-related indicators, but these traits are subjective, discrete, and often detect injury only after visible symptoms have developed. To address these limitations, we developed an integrated hydroponic phenotyping framework that combines standardized salt treatment, RGB imaging, spectral sensing, deep-learning-based tissue segmentation, quantitative trait extraction, and genome-wide association analysis for rice seedling salt-tolerance evaluation. The framework uses a controlled hydroponic cultivation and imaging system to acquire complementary structural, color, and physiological information from rice seedlings. Spectral reflectance across the visible and near-infrared regions was dynamically collected before salt treatment and at multiple time points after treatment. Spectral response analysis identified salt-sensitive wavelengths and vegetation indices associated with subsequent seedling injury and survival. These spectral traits captured physiological changes preceding or accompanying visible symptoms, while their discriminatory ability generally increased with the duration of salt stress, providing a basis for the early and dynamic evaluation of salt tolerance. For RGB phenotyping, we constructed the Rice Seedling-Salt RGB Dataset (RSSD), in which green shoot tissue, yellow shoot tissue, roots, and background were annotated at the pixel level. The ELMERF network integrates a Mix Transformer semantic branch, an Edge-guided LoG-CNN branch, and an Edge-aware Relational Fusion Head to improve the separation of slender roots, green shoots, and gradually yellowing shoot tissues. On the RSSD-A test set, ELMERF achieved a mean Intersection over Union of 51.4% and a mean accuracy of 89.5%, outperforming nine representative semantic segmentation models.
