火力与指挥控制2025，Vol.50Issue(4)：141-149,9.DOI:10.3969/j.issn.1002-0640.2025.04.020
基于YOLOv7-RSobb的遥感图像定向包围框目标检测
Target Detection of Oriented Bounding Boxes in Remote Sensing Images Based on YOLOv7-RSobb
摘要
Abstract
To address the issues of the horizontal bounding box detection in remote sensing image is easy to introduce a lot of background noise for the target with a large ratio of length to width and cannot express the direction information of the target,the YOLOv7-RSobb algorithm is proposed.Based on YOLOv7,the target detection of oriented bounding boxes is realized,PoA and EoE problems are solved,GS-ELAN module is proposed,and Wise-IoU loss function is used to improve mAP.The results show that the mAP of the proposed algorithm on DOTA and DIOR datasets reaches 73.4%and 70.3%,which is significantly improved compared with other mainstream algorithms.关键词
遥感图像/目标检测/环形平滑标签/YOLOv7-RSobb/GS-ELAN/Wise-IoUKey words
remote sensing image/target detection/circular smooth label/YOLOv7-RSobb/GS-ELAN/Wise-IoU分类
信息技术与安全科学引用本文复制引用
梁琦,杨晓文,巩青歌,曹亚明..基于YOLOv7-RSobb的遥感图像定向包围框目标检测[J].火力与指挥控制,2025,50(4):141-149,9.基金项目
山西省自然科学基金资助项目 （）
