Lightweight multi-view collaborative image recognition method of unmanned cluster

• Main Authors: ZHANG BO, XIE ZHIDONG, ZHENG JIANCHAO, PENG JINLIN
• Format: Patent
• Language: Chinese; English
• Published: 25.06.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention provides a lightweight multi-view collaborative image recognition method for an unmanned cluster, and the method comprises the following steps: step 10, constructing a multi-view recognition unmanned cluster system which comprises a head node unmanned aerial vehicle and a plurality of leaf node unmanned aerial vehicles, and the plurality of leaf node unmanned aerial vehicles communicate with the head node unmanned aerial vehicle; the plurality of leaf node unmanned aerial vehicles are distributed in different directions of the identification area; 20, the leaf node unmanned aerial vehicle collects a single-view-angle image of the recognition object, extracts single-view-angle image features and transmits the single-view-angle image features to the head node unmanned aerial vehicle; step 30, the head node unmanned aerial vehicle associates and aggregates the single-view-angle image features sent by all leaf node unmanned aerial vehicles to obtain multi-view-angle image features; and performing id