河北师范大学学报—

期刊信息

刊名：河北师范大学学报（自然科学版）Journal of Hebei Normal University (Natural Science)
主办：河北师范大学
ISSN： 1000-5854
CN： 13-1061/N
中国科技核心期刊
中国期刊方阵入选期刊
中国高校优秀科技期刊
华北优秀期刊
河北省优秀科技期刊

协同采编系统

友情链接

DNEA-YOLO:一种融合双重归一化高效注意力的遥感目标检测模型

刘斌¹

李晓帅²

叶刚¹

梁超¹

胡成玺²，邱磊²

(1.国家能源集团宁夏煤业有限责任公司羊场湾煤矿,宁夏银川 750411; 2.中国矿业大学（北京）人工智能学院，北京 100083)
DOI： 10.13763/j.cnki.jhebnu.nse.202601012

DNEA-YOLO:a remote sensing object detection model integrating dual normalization and efficient attention

Liu Bin¹,Li Xiaoshuai²,Ye Gang¹,Liang Chao¹,Hu Chengxi²,Qiu Lei²

PDF
下载
HTML

摘要/Abstract

摘要：

物联网数据和遥感的融合极大地扩展了遥感领域的目标检测能力，然而遥感目标检测仍面临复杂背景干扰、密集小目标易漏检、模型难以兼顾精度与边缘部署等问题.为此，提出双重归一化高效注意力YOLO（dual normalization efficient attention YOLO，DNEA-YOLO）模型，在满足边缘部署需求的同时提升检测精度.在模型设计中，提出双重归一化高效注意力机制，降低计算开销并增强特征选择能力；针对复杂场景特征提取不足问题，将大核选择模块（large selective kernel，LSK）嵌入C3K2模块的特征融合分支，形成C3K2-LSK集成结构，提升复杂背景下的特征表达能力；针对遥感图像密集目标易漏检问题，采用软非极大值抑制（soft non-maximum suppre-ssion，SOFT-NMS）优化后处理策略，进一步提升目标识别准确性.在SIMD公开数据集上的实验表明，DNEA-YOLO参数量为9.97 M，浮点运算量为25.2 GFlops，mAP较基线模型YOLO11s提升2.8 %.通过与多个先进模型进行对比实验，结果充分验证了DNEA-YOLO在遥感图像目标检测任务中的有效性与优越性.

Abstract：

The integration of IoT data and remote sensing has greatly expanded the object detection capability in the field of remote sensing.However,remote sensing object detection still faces problems such as complex background interference,easy missed detection of dense small targets,and difficulty in balancing accuracy and edge deployment.Therefore,a dual normalization efficient attention YOLO(DNEA-YOLO) model is proposed to improve the detection accuracy while meeting the requirements of edge deployment.In the model design,the dual normalization efficient attention mechanism is proposed to reduce the computational overhead and enhance the feature selection ability.Aiming at the problem of insufficient feature extraction in complex scenes,the large selective kernel(LSK) is embedded into the feature fusion branch of the C3K2 module to form a C3K2-LSK integrated structure to improve the feature expression ability in complex backgrounds.Aiming at the problem of missed detection of dense targets in remote sensing images,soft non-maximum suppression(SOFT-NMS) is used to optimize the post-processing strategy to further improve the accuracy of target recognition.Experiments on the SIMD public dataset show that the DNEA-YOLO parameter is 9.97 M,the floating-point operation is 25.2 GFlops,and the mAP is 2.8 % higher than the baseline model YOLO11s.The comparison results with several advanced models verify the effectiveness and superiority of DNEA-YOLO in remote sensing image object detection tasks.

关键词

关键词： 目标检测；遥感图像；注意力机制；边缘部署； YOLO

Key words： object detection ; remote sensing image ; attention mechanism ; edge deployment ; YOLO

参考文献 28

[1] Ahmed I,Ahmad M,Chehri A,et al.IoT enabled deep learning based framework for multiple object detection in remote sensing images[J].Remote Sensing,2022,14(16):4107.DOI:10.3390/rs14164107.
[2] Xie F,Luo H,Li S Q,et al.Using clean energy satellites to interpret imagery:a satellite IoT oriented lightweight object detection framework for SAR ship detection[J].Sustainability,2022,14(15):9277.DOI:10.3390/su14159277.
[3] Lu A Q,Cheng Y,Hu Y B,et al.Satellite-terrestrial collaborative object detection via task-inspired framework[J].IEEE Internet of Things Journal,2023,10(23):20528-20544.DOI:10.1109/jiot.2023.3287973.
[4] Deng Z P,Sun H,Zhou S L,et al.Multi-scale object detection in remote sensing imagery with convolutional neural networks[J].ISPRS Journal of Photogrammetry and Remote Sensing,2018,145:3-22.DOI:10.1016/j.isprsjprs.2018.04.003.
[5] 涂兵,张晓飞,张国云,等.递归滤波与KNN的高光谱遥感图像分类方法[J].国土资源遥感,2019,31(1):22-32.DOI:10.6046/gtzyyg.2019.01.04. Tu Bing,Zhang Xiaofei,Zhang Guoyun,et al.Hyperspectral image classification via recursive filtering and KNN[J].Remote Sensing for Land & Resources,2019,31(1):22-32.
[6] Cheng G,Si Y J,Hong H L,et al.Cross-scale feature fusion for object detection in optical remote sensing images[J].IEEE Geoscience and Remote Sensing Letters,2021,18(3):431-435.DOI:10.1109/LGRS.2020.2975541.
[7] Liu Y F,Li Q,Yuan Y,et al.ABNet:adaptive balanced network for multiscale object detection in remote sensing imagery[J].IEEE Transactions on Geoscience and Remote Sensing,2022,60:5614914.DOI:10.1109/TGRS.2021.3133956.
[8] 周莲,何楚,汪鼎文,等.光学遥感图像的小样本目标检测[J].遥感学报,2024,28(7):1693-1701.DOI:10.11834/jrs.20243209. Zhou Lian,He Chu,Wang Dingwen,et al.Few-shot object detection in optical remote sensing images[J].National Remote Sensing Bulletin,2024,28(7):1693-1701.
[9] 刘浩然,严天笑,朱月琴,等.基于一种改进YOLOv7算法模型的滑坡识别研究:以四川省白格地区为例[J].自然资源遥感,2025,37(4):48-57.DOI:10.6046/zrzyyg.2024110. Liu Haoran,Yan Tianxiao,Zhu Yueqin,et al.Landslide identification based on an improved YOLOv7 model:a case study of the Baige area[J].Remote Sensing for Natural Resources,2025,37(4):48-57.
[10] 李愿,付辉,刘浩志.双重注意力下的多尺度残差遥感图像去雾网络[J].自然资源遥感,2025,37(4):31-39.DOI:10.6046/zrzyyg.2024154. Li Yuan,Fu Hui,Liu Haozhi.Multi-scale residual dehazing network for remote sensing images based on dual attention[J].Remote Sensing for Natural Resources,2025,37(4):31-39.
[11] Hu J,Shen L,Albanie S,et al.Squeeze-and-excitation networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2020,42(8):2011-2023.DOI:10.1109/TPAMI.2019.2913372.
[12] Wang Q L,Wu B G,Zhu P F,et al.ECA-Net:efficient channel attention for deep convolutional neural networks[C]//IEEE.2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Seattle:IEEE,2020:11531-11539.DOI:10.1109/cvpr42600.2020.01155.
[13] Yang L X,Zhang R Y,Li L D,et al.SimAM:A simple,parameter-free attention module for convolutional neural networks[C]//Marina M,Tong Z.Proceedings of the 38th International Conference on Machine Learning.New York:PMLR,2021:11863-11874.DOI:10.1109/ICASSP43922.2022.9746294.
[14] Liu Y,Shao Z,Hoffmann N.Global attention mechanism:retain information to enhance channel-spatial interactions[PP/OL].arXiv(2021-12-10)[2025-11-06].https://arxiv.org/abs/2112.05561.
[15] Liu Y C,Shao Z R,Teng Y Y,et al.NAM:normalization-based attention module[PP/OL].arXiv(2021-11-24)[2025-06-07].https://doi.org/10.48550/arXiv.2111.12419.
[16] Xia Z F,Pan X R,Song S J,et al.Vision Transformer with deformable attention[C]//IEEE.2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).New Orleans:IEEE,2022:4784-4793.DOI:10.1109/cvpr52688.2022.00475.
[17] Shao H,Zeng Q S,Hou Q B,et al.MCANet:medical image segmentation with multi-scale cross-axis attention[J].Machine Intelligence Research,2025,22(3):437-451.DOI:10.1007/s11633-025-1552-6.
[18] Xu S B,Zheng S C,Xu W H,et al.HCF-Net:hierarchical context fusion network for infrared small object detection[C]//IEEE.2024 IEEE International Conference on Multimedia and Expo(ICME).Niagara Falls:IEEE,2024:1-6.DOI:10.1109/icme57554.2024.10687431.
[19] Balakrishna S,Yadao S,Solanki V K.Enhancing YOLOv11 for real-time object detection:advanced architectures and edge-optimized training pipeline[C]//Vijender K S,Tran D T,Pradeep K,et al.Proceedings of the Ninth International Conference on Research in Intelligent Computing in Engineering.Warszawa:Polish Information Processing Society,2025:89-96.DOI:10.15439/2024R115.
[20] Li Y X,Hou Q B,Zheng Z H,et al.Large selective kernel network for remote sensing object detection[C]//IEEE.2023 IEEE/CVF International Conference on Computer Vision(ICCV).Paris:IEEE,2023:16748-16759.DOI:10.1109/ICCV51070.2023.01540.
[21] Bodla N,Singh B,Chellappa R,et al.Soft-NMS:improving object detection with one line of code[C]//IEEE.2017 IEEE International Conference on Computer Vision(ICCV).Venice:IEEE,2017:5562-5570.DOI:10.1109/iccv.2017.593.
[22] Haroon M,Shahzad M,Fraz M M.Multisized object detection using spaceborne optical imagery[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2020,13:3032-3046.DOI:10.1109/JSTARS.2020.3000317.
[23] Misra D,Nalamada T,Arasanipalai A U,et al.Rotate to attend:convolutional triplet attention module[C]//IEEE.2021 IEEE Winter Conference on Applications of Computer Vision(WACV).Waikoloa:IEEE,2021:3138-3147.DOI:10.1109/wacv48630.2021.00318.
[24] Zhang J N,Li X T,Li J,et al.Rethinking mobile block for efficient attention-based models[C]//IEEE.2023 IEEE/CVF International Conference on Computer Vision(ICCV).Paris:IEEE,2023:1389-1400.DOI:10.1109/iccv51070.2023.00134.
[25] Dai W,Liu R,Wu Z X,et al.Exploiting scale-variant attention for segmenting small medical objects[J].IEEE Transactions on Neural Networks and Learning Systems,2026,37(7):3161-3175.DOI:10.1109/tnnls.2025.3645355.
[26] Wan D H,Lu R S,Shen S Y,et al.Mixed local channel attention for object detection[J].Engineering Applications of Artificial Intelligence,2023,123:106442.DOI:10.1016/j.engappai.2023.106442.
[27] Zhang H,Ma Z,Li X.RS-DETR:an improved remote sensing object detection model based on RT-DETR[J].Applied Sciences,2024,14(22):10331.DOI:10.3390/app142210331.
[28] Liu Y L,Yang D G,Song T T,et al.YOLO-SSP:an object detection model based on pyramid spatial attention and improved downsampling strategy for remote sensing images[J].The Visual Computer,2025,41(3):1467-1484.DOI:10.1007/s00371-024-03434-y.