Detection method for aerial photography of dense small targets by unmanned aerial vehicle

The invention provides a detection method for aerial photography of dense small targets by an unmanned aerial vehicle. The method comprises the following steps: firstly, constructing a lightweight backbone network CSPDarknet-tiny on the basis of a YOLOv7 model, so that the down-sampling rate is reduced, and more semantic information and detail features are reserved; secondly, a multi-head attention mechanism MHSA is introduced into the neck network, so that the interference of irrelevant information caused by a complex background is effectively relieved, the network is helped to pay more attention to feature information extraction of the small target, and the detection precision of the small target is improved; and finally, introducing NWD Loss for the sensitivity of the IOU Loss to the position difference of the small target, and combining the IOU Loss and the NWD Loss through a certain weight proportion to obviously improve the detection precision of the small target. The small target detection problem unde