This invention belongs to the field of
plant scanning AI glasses technology, and particularly relates to a
plant scanning method and
plant scanning AI glasses, including an AI glasses terminal and a
smartphone APP. The AI glasses terminal integrates a side-end real-time visual acquisition and
processing module, an optical display module, and an environmental adaptive
perception unit for collecting environmental parameters of the
work site. The side-end real-time visual acquisition and
processing module integrates a 4K global
shutter camera and an edge AI computing unit. The edge AI computing unit deploys a lightweight
deep learning model and stores a local
feature vector library, which can complete plant target identification and positioning in real time without a
network connection. It can flexibly adjust the confidence judgment threshold value according to the operator's experience,
regional differences, and real-time weather conditions, and simultaneously achieves high species fine-grained identification accuracy and full environmental element adaptive capability on an extremely lightweight
edge device.