火力与指挥控制2025，Vol.50Issue(7)：42-49,54,9.DOI:10.3969/j.issn.1002-0640.2025.07.006
基于尺度自适应YOLOv8的遥感目标检测算法
Remote Sensing Target Detection Algorithm Based on Scale-adaptive YOLOv8
摘要
Abstract
Aiming at the detection difficulties such as low resolution of military remote sensing target detection images,different target scales and target aliasing,etc.YOLO-MRO target detection algorithm is proposed.First,the adaptive convolution module is constructed to improve the detection effect of different scale targets.Secondly,wavelet pooling convolution is designed based on wavelet pooling,the ability of the algorithm to resist aliasing is enhanced.Finally,hybrid attention is introduced to reduce the computational burden of the algorithm.The experimental results show that the improved algorithm improves the accuracy by 6.7%compared to the benchmark algorithm on the self-built dataset RSMOD,reaching 81.1%,which is able to meet the detection task of military remote sensing targets.关键词
军事目标识别/光学遥感/自适应融合/混合注意力Key words
military target recognition/optical remote sensing/adaptive fusion/hybrid attention分类
军事科技引用本文复制引用
张灿,李志刚,袁一东,李莹琦..基于尺度自适应YOLOv8的遥感目标检测算法[J].火力与指挥控制,2025,50(7):42-49,54,9.基金项目
国家自然科学基金区域创新发展联合基金重点资助项目(U21A20114) （U21A20114）
河北省军民融合发展专项基金资助项目(SJMYF2022X05) （SJMYF2022X05）
