YOLOv5-SATC:用于遥感图像目标检测的网络

YOLOv5-SATC:A Network for Object Detection in Remote Sensing Images

无人机(Unmanned Aerial Vehicle,UAV)遥感图像具有成像距离远、分辨率高和目标密集等特点。本文旨在开发一种基于深度神经网络和注意力机制的学习模型,以改善无人机遥感视角下微小目标检测的性能。本研究改进了YOLOv5,并提出了一种名为YOLOv5-SATC的新无人机遥感图像微小目标检测算法。实验结果表明,该算法在NWPU数据集上优于其他主流目标检测方法。

Unmanned Aerial Vehicle(UAV)remote sensing images have the characteristics of long imaging distances,high resolutions,and dense target distribution.This paper aims to develop a learning model based on deep neural networks and attention mechanisms to improve the performance of small object detection in UAV remote sensing perspectives.This study improve YOLOv5 and propose a new algorithm for small object detection in UAV remote sensing images,named YOLOv5-SATC.Finally,we conducted a series of comparative experiments on the NWPU dataset,and those results demonstrate that our proposed algorithm outperforms other mainstream object detection methods on the NWPU dataset.