Abstract: A lightweight target detection algorithm for UAV aerial images is proposed to address the problems of large model size and low detection accuracy in traditional UAV aerial image target detection.Based on YOLOv5s,a small target detection layer is added,and a omni-dimensional dynamic convolution is used to replace the ordinary convolution,which reduces the number of parameters.Using cross-layer and cross-scale weighted feature fusion,and introducing FasterNet module,the feature extraction capability is strengthened.A dynamic label assignment strategy is used to significantly improve the detection accuracy.The experimental results show that the proposed algorithm outperforms the original YOLOv5s algorithm in terms of accuracy and model volume,and can more efficiently accomplish the task of target detection in UAV aerial images.