The invention provides a method for generating a global operation path of
agricultural machinery in a tea-oil garden based on unmanned aerial vehicle vision, which comprises the following steps of: firstly, acquiring images by using an unmanned aerial vehicle to construct a
data set, and performing improvement based on a YOLOv11 model: embedding an EMA attention mechanism to enhance feature focusing capability, and adopting a SlideLoss
loss function to optimize multi-scale crown positioning precision; then, the trained model is used for detecting a tree crown and extracting a central point of the tree crown; and finally, performing clustering and straight
line fitting based on central point distribution to extract a navigation baseline, and further generating a smooth global operation path shaped like a Chinese character'gong '. According to the method, only the
consumer-level unmanned aerial vehicle and the RGB camera are needed, so that the cost is low; through a new normal form of planning through detection, complex three-dimensional reconstruction is avoided, and the process is efficient; the detection precision on a self-built
data set is excellent; the generated path is highly similar to the GNSS reference path, and a solution with high
cost performance is provided for autonomous operation of the
agricultural machinery in the
camellia oleifera garden.