电子学报2023，Vol.51Issue(11)：3238-3247,10.DOI:10.12263/DZXB.20220612
面向遥感目标检测的无锚框Transformer算法
Anchor-Free Transformer Algorithm for Aerial Remote Sensing Target Detection
摘要
Abstract
Aerial remote sensing image targets have the characteristics of multi-directional arrangement,small,and dense.The rotating target detection algorithm based on deep learning has the problem of poor detection accuracy.To solve this problem,the article proposes a novel anchor-free Transformer algorithm for aerial remote sensing target detection.Firstly,hierarchical Transformer is used to collect feature information of different resolutions to improve the range of fea-ture information collection.Secondly,a new feedforward network(Spacial-FeedForward Neural network,SFFN)is con-structed.SFFN combines the local space characteristics of 3×3 depth separable convolution with the global channel charac-teristics of multi-layer perceptron(MLP)to solve the shortcomings of feed forward neural network(FFN)in local space modeling.Finally,an anchor-free detector is built based on SFFN architecture,and the regression problem of prediction frame is divided into horizontal frame and rotating frame,which alleviates the loss discontinuity problem of rotating frame.The test results on DOTA dataset show that the average accuracy of this method has reached 75.83%,respectively,while achieving 92.47%of 5 small targets on NWPU VHR-10 dataset,which is more competitive in remote sensing target detec-tion accuracy.关键词
遥感图像/目标检测/Transformer算法/无锚框检测器Key words
remote sensing image/target detection/Transformer algorithm/anchor-free detector分类
信息技术与安全科学引用本文复制引用
喻九阳,胡天豪,戴耀南,张德安,夏文凤..面向遥感目标检测的无锚框Transformer算法[J].电子学报,2023,51(11):3238-3247,10.基金项目
湖北省重点研发计划(No.2020BAB030) （No.2020BAB030）
湖北省自然科学基金(No.2023AFC010)Hubei Provincial Key R&D Project(No.2020BAB030) （No.2023AFC010）
Natural Science Foundation of Hubei Province(No.2023AFC010) （No.2023AFC010）
