A lightweight target recognition algorithm for UAV

Aiming at the accuracy and real-time requirements of UAV scene, this paper proposes a lightweight target recognition algorithm based on multi mechanism collaborative optimization. Firstly, a new PCE module is designed for the feature extraction part, which uses partial convolution to reduce the amount of calculation, and introduces an efficient multi-scale attention mechanism to enhance the feature extraction ability of the model; Secondly, a lightweight detection head architecture based on parameter sharing mechanism and structural heavy parameterization technology is proposed, which not only greatly reduces the amount of parameters and calculation, but also effectively maintains the target detection accuracy of the model in UAV application scenarios; In addition, the WIOU loss function is introduced to reduce the prediction bias of the model and improve the generalization ability of the model. The experiment is carried out on the UAV public data set EVD4UAV, and the results show that the proposed algorithm achieves a good balance in real-time and accuracy in the UAV application scenario.