Precision variable-rate nitrogen topdressing application for individual cabbage using unmanned aerial vehicle-derived maps and a fuzzy decision support system

In this study, we developed and evaluated a variable-rate nitrogen (VRN) fertilization strategy for cabbage by integrating soil measurements, unmanned aerial vehicle (UAV) derived vegetation indices, and yield data. A fuzzy gamma overlay approach was used to generate the first topdressing (ToPD) nitrogen prescription map by combining soil nitrogen, cation exchange capacity, nitrogen reflectance index, and estimated yield into a standardized decision surface. For the second ToPD, UAV-based vegetation and growth parameters – including the normalized difference vegetation index (NDVI), green NDVI (GNDVI), and normalized difference red edge index (NDRE) – were used to refine the spatial fertilizer application recommendations. Fuzzy membership functions enabled the harmonization of heterogeneous datasets and enhanced the robustness of prescription maps. Field application results demonstrated that the first ToPD, applied five weeks after transplanting, did not significantly influence aboveground biomass and head fresh weight at the time of harvesting. The second ToPD, applied three weeks after the first ToPD, significantly increased head fresh weight, aboveground biomass, and head diameter. The yield and growth uniformity in 2025 confirmed substantial improvements, with cabbage head fresh weight in the range of 1.4–4.1 kg in both plots. These findings underscore that the timing of fertilizer application is an important factor influencing cabbage growth.