Longitudinal RGB and Ultraviolet-Excited Tomato Imaging Dataset During 18-Day Storage

This dataset supports the research article entitled “Longitudinal Visible and Ultraviolet-Excited Imaging of Tomato Changes During Storage.” It contains longitudinal image-derived data obtained from an untreated red-stage tomato monitored over 18 consecutive days under ambient storage conditions. The dataset was developed for non-destructive characterization of storage-related changes using complementary RGB and ultraviolet-excited imaging. The RGB dataset contains eight appearance descriptors: mean red-channel intensity (MeanR), mean green-channel intensity (MeanG), mean blue-channel intensity (MeanB), colorfulness, segmented image area, RGB entropy, smoothness, and image contrast. The ultraviolet-excited dataset contains four Gray-Level Co-Occurrence Matrix (GLCM) texture descriptors—contrast, correlation, energy, and homogeneity—together with grayscale image entropy. GLCMs were generated using eight gray levels over the 8-bit grayscale intensity range, with a one-pixel horizontal offset [0 1] corresponding to 0°, Symmetric = false, and no directional averaging. Contextual environmental measurements recorded during the 18-day storage period are also provided, including temperature, relative humidity, equivalent CO₂ (eCO₂), and total volatile organic compounds (TVOC). These environmental variables are supplied as contextual storage metadata and were not used as predictors in the RGB–UV cross-modal analysis. The accompanying spreadsheet includes the complete 18-day numerical descriptors, normalized longitudinal trajectories, descriptive Spearman rank associations with storage day, cross-modal Spearman associations, a data dictionary, and quality-control checks supporting the values reported in the manuscript. This dataset represents repeated longitudinal observations of a single biological specimen and should not be interpreted as 18 independent biological replicates. The image-derived descriptors characterize changes in visible appearance and ultraviolet-excited spatial texture and should not be interpreted as direct measurements of physiological freshness or physicochemical quality. Segmented image area is expressed in pixels and represents two-dimensional image-space extent rather than calibrated physical surface area. The dataset is intended to support reproducibility, methodological transparency, and further research on interpretable computer-vision approaches for longitudinal postharvest imaging.