 Abstract YRL of 5 is one of the most popular object detection algorithms, which can be divided into multiple series depending on the control of network depth and width. The paper proposes a lightweight aerial image object detection algorithm, LAI YRL of 5s, based on improvements to the original YRL of 5 algorithm, including replacing the minimum detection head with the maximum detection head and introducing a new feature fusion method, deep feature map cross-path fusion network, DFMC-PFM. Additionally, a new module based on VOVnet is designed to enhance the feature extraction capabilities of the backbone network. Finally, the paper utilizes the idea of ShuffleNet V2 to make the network lighter while maintaining its accuracy. On the Vistrone 2019 dataset, the detection accuracy of LAI YRL of 5s is 8.3% higher than that of the original algorithm. This article was offered by Lixia Deng, Linyun Bai, Hongchuan Li, and others. We are article.tv, links in the description below.