Remote Sensing Object Detection Based on Convolution and Swin Transformer

被引:14
|
作者
Jiang, Xuzhao [1 ]
Wu, Yonghong [1 ]
机构
[1] Wuhan Univ Technol, Dept Stat, Wuhan 430070, Peoples R China
基金
中国国家自然科学基金;
关键词
Object detection; Feature extraction; Transformers; Remote sensing; Prediction algorithms; Detection algorithms; Classification algorithms; Remote sensing images; object detection; attention mechanism; swin transformer; multi-scale features;
D O I
10.1109/ACCESS.2023.3267435
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Remote sensing object detection is an essential task for surveying the earth. It is challenging for the target detection algorithm in natural scenes to obtain satisfactory detection results in remote sensing images. In this paper, the RAST-YOLO (You only look once with Regin Attention and Swin Transformer) algorithm is proposed to address the problems of remote sensing object detection, such as significant differences in target scales, complex backgrounds, and tightly arranged small-size targets. To increase the information interaction range of the feature map, make full use of the background information of the object, and improve the detection accuracy of the object with a complex background, the Regin Attention (RA) mechanism combined with Swin Transformer as the backbone is proposed to extract features. To improve the detection accuracy of small objects, the C3D module is used to fuse deep and shallow semantic information and optimize the multi-scale problem of remote sensing targets. To evaluate the performance of RAST-YOLO, extensive experiments are performed on DIOR and TGRS-HRRSD datasets. The experimental results show that RAST achieves state-of-the-art detection accuracy with high efficiency and robustness. Specifically, compared with the baseline network, the mean average precision (mAP) of detection results is improved by 5% and 2.3% on DIOR and TGRS-HRRSD datasets, respectively, which demonstrates RAST-YOLO is effective and superior. Moreover, the lightweight structure of RAST-YOLO can ensure the real-time detection speed and obtain excellent detection results.
引用
收藏
页码:38643 / 38656
页数:14
相关论文
共 50 条
  • [21] M-Swin: Transformer-Based Multiscale Feature Fusion Change Detection Network Within Cropland for Remote Sensing Images
    Pan, Jun
    Bai, Yuchuan
    Shu, Qidi
    Zhang, Zhuoer
    Hu, Jiarui
    Wang, Mi
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2024, 62 : 1 - 16
  • [22] HIERARCHICAL REGION BASED CONVOLUTION NEURAL NETWORK FOR MULTISCALE OBJECT DETECTION IN REMOTE SENSING IMAGES
    Li, Qingpeng
    Mou, Lichao
    Jiang, Kaiyu
    Liu, Qingjie
    Wang, Yunhong
    Zhu, Xiao Xiang
    IGARSS 2018 - 2018 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM, 2018, : 4355 - 4358
  • [23] Multi-scale Feature Fusion Object Detection Based on Swin Transformer
    Zhang, Ying
    Wu, Lin
    Deng, Huaxuan
    Hu, Jun
    Li, Xifan
    39TH YOUTH ACADEMIC ANNUAL CONFERENCE OF CHINESE ASSOCIATION OF AUTOMATION, YAC 2024, 2024, : 1982 - 1987
  • [24] DEST: Difference enhanced-Swin Transformer for remote sensing change detection
    Wang, Xin
    Zeng, Zeyang
    Li, Li
    REMOTE SENSING LETTERS, 2024, 15 (12) : 1229 - 1238
  • [25] Adaptive Spatial Tokenization Transformer for Salient Object Detection in Optical Remote Sensing Images
    Gao, Lina
    Liu, Bing
    Fu, Ping
    Xu, Mingzhu
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2023, 61
  • [26] Remote Sensing Image Fusion Method Based on Improved Swin Transformer
    Li Zitong
    Zhao Jiankang
    Xu Jingran
    Long Haihui
    Liu Chuanqi
    ACTA PHOTONICA SINICA, 2023, 52 (11)
  • [27] DETR Novel Small Target Detection Algorithm Based on Swin Transformer
    Xu, Fengchang
    Alfred, Rayner
    Pailus, Rayner Henry
    Lyu, Ge
    Du, Shifeng
    Chew, Jackel Vui Lung
    Li, Guozhang
    Wang, Xinliang
    IEEE ACCESS, 2024, 12 : 115838 - 115852
  • [28] Transformer with Transfer CNN for Remote-Sensing-Image Object Detection
    Li, Qingyun
    Chen, Yushi
    Zeng, Ying
    REMOTE SENSING, 2022, 14 (04)
  • [29] An Empirical Study of the Convolution Neural Networks Based Detection on Object With Ambiguous Boundary in Remote Sensing Imagery-A Case of Potential Loess Landslide
    Yao, Guangle
    Zhou, Wenlong
    Liu, Mingzhe
    Xu, Qiang
    Wang, Honghui
    Li, Jun
    Ju, Yuanzhen
    IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2022, 15 : 323 - 338
  • [30] Object Detection for Remote Sensing Based on the Enhanced YOLOv8 With WBiFPN
    Shen, Lingyun
    Lang, Baihe
    Song, Zhengxun
    IEEE ACCESS, 2024, 12 : 158239 - 158257