Detection of Nitrogen and Water-Deficit Stress in Single-plant and Plug-tray Units of Tomato Seedlings using Spectral Imaging

Hyperspectral and multispectral imaging data have recently been used to support research that evaluated specific abiotic stress levels in single-plant units. However, for practical applications in a seedling nursery, assessing abiotic stress levels in plug-tray units is necessary. Therefore, this study aimed to develop techniques for identifying nitrogen and water-deficit-stress levels in tomato seedlings in both single-plant and plug-tray units using spectral imaging. Nitrogen-stress treatments included five levels, as well as a control, whereas water-deficit-stress treatments included two levels: a control and a root-zone drying treatment. Spectral reflectance and vegetation indices were calculated using spectral data for different wavelengths obtained from multispectral or hyperspectral cameras. Correlations between spectral image values (reflectance and vegetation indices) and stress-related substances (chlorophyll, proline, and total nitrogen content) were analyzed. It was possible to evaluate by means of multispectral and hyperspectral imaging the nitrogen-stress levels corresponding to the treatments with a seedling nitrogen content of less than 0.89% as a result of restricting nitrogen nutrients for 26 days. Vegetation indices (normalized difference vegetation index [NDVI], green normalized difference vegetation index [GNDVI], and green chlorophyll index [CIgreen]) calculated from multispectral imaging indicated that nitrogen stress could be detected in the treatment that excluded nitrogen for 16 days, with a nitrogen content of 1.30% or less. For water-deficit stress, it was not possible to detect stress levels using multispectral or hyperspectral data within the 1,000 nm wavelength range. Thus, while nitrogen-stress levels can be practically detected at the plug-tray unit level by means of a multispectral or hyperspectral analysis, detecting water-deficit-stress levels using near-infrared wavelengths below 1,000 nm is difficult. This study demonstrated that multispectral or hyperspectral imaging can be used as a practical means of assessing nitrogen-stress levels in single plants and plug-tray units.