Unmanned vehicle dynamic selection method for dam intelligent collaborative inspection

The invention discloses an unmanned vehicle dynamic selection method for dam intelligent collaborative inspection, which performs corresponding inspection tasks in a multi-node collaborative manner aiming at the isomerism of edge equipment based on federated learning in a real dam reservoir area environment, and comprises the following steps: constructing a local calculation efficiency evaluation standard of nodes; grouping according to the representation of the computing power of the node client; selecting a weight through local data representativeness evaluation construction in the group, selecting a client with the most data representativeness to participate in local training, and fusing results of each group to complete client selection; and after aggregation of each iteration round is finished, updating the local computing efficiency of the client according to the training time of the latest round, and repeating the steps to dynamically update a client selection result. According to the method, the parti