Vehicle re-identification method based on grouping aggregation attention and local relation

The invention relates to the technical field of vehicle re-identification in computer vision, in particular to a vehicle re-identification method based on grouping aggregation attention and local relation, which is realized based on a grouping aggregation attention mechanism and a local relation module, feature enhancement is independently carried out on each group, then cross-group local information interaction is realized by utilizing an aggregation mode, and attention weights of channels are effectively inferred by utilizing different global relationships. Secondly, an attention enhancement constraint is applied to the attention map, and the constraint can adaptively adjust the weight value of a channel, so that the discriminative information is further enhanced, the noise information is inhibited, and the ability of the network to identify the same vehicle is improved; the local relation module aims at mining the relation between more valuable parts so as to distinguish the vehicles with different identit