The invention discloses an
orchard operation
robot global positioning method and
system based on a large
language model-multi-metric ICP (
Inductively Coupled Plasma). The method comprises the following steps: firstly, constructing a global semantic priori map containing
structured text description and macroscopic geometric feature vectors; then real-time
point cloud data are collected, and text description and geometric feature vectors of a current scene are generated through semantic segmentation; a coarse-to-fine two-stage registration strategy is adopted, in the coarse registration stage, features are screened through information entropy, an optimal candidate area is screened in combination with adaptive fusion scores of
semantic matching degree and geometric feature similarity, and a coarse registration
pose is obtained through an NDT
algorithm; in the fine registration stage, a multi-metric ICP optimization objective function fusing four residual terms of distance, feature, strength and
semantics is constructed, the coarse registration
pose is used as an initial value, the weight of each metric term is dynamically adjusted based on a large
language model, and optimal
pose transformation is solved; and obtaining a global positioning result of the
robot based on the optimal pose transformation. According to the invention, the positioning precision and robustness are improved.