Computer Engineering and Applications ›› 2023, Vol. 59 ›› Issue (9): 207-214.DOI: 10.3778/j.issn.1002-8331.2209-0119

• Graphics and Image Processing • Previous Articles     Next Articles

Target Detection Algorithm of Remote Sensing Image Based on Improved YOLOv5

LI Kunya, OU Ou, LIU Guangbin, YU Zefeng, LI Lin   

  1. College of Computer & Internet Security(College of Oxford Brooks), Chengdu University of Technology, Chengdu 610051, China
  • Online:2023-05-01 Published:2023-05-01

改进YOLOv5的遥感图像目标检测算法

李坤亚,欧鸥,刘广滨,于泽峰,李林   

  1. 成都理工大学 计算机与网络安全学院(牛津布鲁克斯学院),成都 610051

Abstract: Aiming at the problems of low target detection accuracy caused by high background complexity, multiple target sizes and too many small targets in remote sensing images, this paper proposes a target detection algorithm of remote sensing image based on improved YOLOv5. The channel-global attention mechanism(CGAM) is introduced into the backbone network to enhance the feature extraction ability of targets at different scales and to suppress the interference of redundant information. The dense upsampling convolution(DUC) module is introduced to expand the low resolution convolution feature maps, which can effectively enhance the fusion effect of different convolution feature maps. The improved algorithm is applied to the open remote sensing data set RSOD, and the average accuracy AP value of the improved YOLOv5 algorithm reaches 78.5%, which is 3.1?percentage points higher than that of the original algorithm. Experimental results show that the improved algorithm can effectively improve the accuracy of remote sensing image target detection.

Key words: remote sensing image, target detection, YOLOv5, channel attention mechanism, upsampling

摘要: 针对遥感图像中背景复杂度高、目标尺寸多样和小目标存在过多所导致的目标检测精度较低的问题,提出一种改进YOLOv5的遥感图像目标检测算法。该算法在主干网络引入通道-全局注意力机制(CGAM)以增强对不同尺度目标的特征提取能力和抑制冗余信息的干扰。引入密集上采样卷积(DUC)模块扩张低分辨率卷积特征图,有效增强不同卷积特征图的融合效果。将改进算法应用于公开遥感数据集RSOD中,改进YOLOv5算法平均精度AP值达到78.5%,较原算法提升了3.1个百分点。实验结果证明,改进后的算法能有效提高遥感图像目标检测精度。

关键词: 遥感图像, 目标检测, YOLOv5, 注意力机制, 上采样