The invention discloses a smart agricultural variable fertilization
control system and method based on image recognition, and relates to the technical field of smart
agriculture, and the method comprises the steps: collecting an
orchard multi-
source image and
laser radar data through an unmanned aerial vehicle, and dividing an
orchard into different
microclimate management units according to topographic features; extracting the
vein mesophyll
color difference ratio of the citrus leaves from the visible light image, correcting the traditional
spectral index, and eliminating the interference of light environment difference on
nutrition diagnosis; based on the corrected
spectral index, establishing a
nitrogen demand model in combination with
terrain and tree body parameters, and generating a single-
plant personalized variable fertilization prescription map; and importing the prescription map into the intelligent
fertilizer applicator, and controlling
fertilizer discharge to execute variable-rate
fertilizer application operation by locating and inquiring the fertilizer application amount in real time. According to the method, the dynamic compensation model based on the
vein and mesophyll
color difference ratio is constructed, the leaf morphological characteristics and the spectral information are cooperatively processed, the traditional
vegetation index is dynamically corrected, and the
spectral response deviation caused by the light environment difference under the complex
terrain can be reduced.