火力与指挥控制2024，Vol.49Issue(4)：24-30,7.DOI:10.3969/j.issn.1002-0640.2024.04.004
基于Ghost-YOLOv5s的SAR图像舰船目标检测
Ship Target Detection in SAR Images Based on Ghost-YOLOv5s
摘要
Abstract
In the ship target detection based on spaceborne synthetic aperture radar(SAR)images,in order to balance the model size and detection accuracy,a ship target detection method Ghost-YOLOv5s based on SAR images of Ghost convolution is proposed.First,Ghost convolution is introduced into the neck of YOLOv5s to reduce the model parameters and to compress model volume.Then,an ef-ficient channel attention mechanism(ECA)is integrated into the C3 module of the neck to highlight the important features so as to maintain high detection performance.Finally,the SIoU loss function is used to replace the original CIoU loss function,so as to reduce the deviation between the predicted box and the real box and to improve the accuracy of the detection algorithm.The experimental results show that compared with YOLOv5s on SSDD remote sensing data set,the improved model reduces the number of model parameters by 6.28%,the model volume by 6.21%,and the detection accuracy reaches 98.21%,achieving the balance between model size and detection accuracy.关键词
合成孔径雷达/深度学习/Ghost卷积/注意力机制Key words
synthetic aperture radar/deep learning/Ghost convolution/attention mechanism分类
信息技术与安全科学引用本文复制引用
张慧敏,黄炜嘉,李锋..基于Ghost-YOLOv5s的SAR图像舰船目标检测[J].火力与指挥控制,2024,49(4):24-30,7.基金项目
国家自然科学基金资助项目(61701416) （61701416）
