Remote sensing small object detection with dynamic perception and cross-modulation

doi:10.11772/j.issn.1001-9081.2025080985

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (9): 2968-2976.DOI: 10.11772/j.issn.1001-9081.2025080985

• Multimedia computing and computer simulation • Previous Articles

Remote sensing small object detection with dynamic perception and cross-modulation

Hua YAO(), Gaoming YANG, Xuelian LI, Kaixuan LU

School of Computer Science and Engineering，Anhui University of Science and Technology，Huainan Anhui 232001，China

Received:2025-08-28 Revised:2025-10-10 Accepted:2025-10-16 Online:2026-09-16 Published:2026-09-10
Contact: Hua YAO
About author:YAO Hua，born in 2001， M. S. candidate. His research interestsinclude object detection.
YANG Gaoming， born in 1974， Ph. D.， professor. His researchinterests include computer vision， digital image processing.
LI Xuelian，born in 2005， M. S. candidate. Her research interestsinclude industrial anomaly detection.
LU Kaixuan， born in 2000， M. S. candidate. His research interestsinclude object detection.

基于动态感知和交叉调制的遥感小目标检测

姚华(), 杨高明, 李雪莲, 陆凯旋

安徽理工大学计算机科学与工程学院，安徽淮南 232001

通讯作者: 姚华
作者简介:姚华（2001—），男，安徽铜陵人，硕士研究生，CCF会员，主要研究方向：目标检测
杨高明（1974—），男，安徽阜阳人，教授，博士，主要研究方向：计算机视觉、数字图像处理
李雪莲（2005—），女，安徽淮南人，硕士研究生，主要研究方向：工业异常检测
陆凯旋（2000—），男，安徽合肥人，硕士研究生，主要研究方向：目标检测。
基金资助:
国家自然科学基金资助项目(52374155);安徽省自然科学基金资助项目(2308085MF218)

Abstract

Abstract:

To address the challenges of severe background interference and insufficient multi-scale feature extraction and fusion for small object detection in remote sensing images， which lead to degraded detection accuracy， a Dynamically Perceptive and Cross-Modulated Network （DPCMNet） was proposed. First， to enhance the model’s perception ability to fine-grained objects， the multi-level Hybrid Feature Aggregation （HFA） module was designed to strengthen the response intensity of small object regions dynamically， thereby enriching the textural and structural features effectively. Second， the Dynamic Modulation Fusion （DMF） module was proposed to introduce a joint channel and spatial modulation mechanism， so as to enable adaptive feature interaction， thereby alleviating information loss during multi-scale feature fusion and improving localization and recognition capabilities for small objects. Finally， based on the DMF module， the Dynamic Modulation Fusion Pyramid （DMFP） was constructed to achieve progressive feature fusion and enhancement through cross-scale association between high-level semantic information and low-level detail compensation. Experimental results demonstrate that DPCMNet achieves significant performance improvements on the Unicorn Small Object Dataset （USOD） and VEhicle Detection in Aerial Imagery （VEDAI） remote sensing small object datasets， with Average Precision （AP） increased by 2.9 and 3.1 percentage points， respectively， compared to the baseline model YOLOv11n.

Key words: remote sensing image, small object detection, context information enhancement, feature interaction, multi-scale fusion

摘要：

针对遥感图像中小目标检测背景干扰严重以及多尺度特征提取和融合不充分导致检测精度下降的问题，提出一种动态感知和交叉调制网络（DPCMNet）。首先，为增强模型对细粒度目标的感知能力，设计多层次的混合特征聚合（HFA）模块，以动态强化小目标区域的响应强度，有效地丰富纹理和结构特征；其次，提出动态调制融合（DMF）模块，引入通道和空间联合调制机制，实现特征间的自适应交互，从而缓解多尺度融合过程中的信息丢失，并提升小目标的定位和识别能力；最后，基于DMF模块构建动态调制融合金字塔（DMFP），通过高层语义信息和浅层细节补偿的跨尺度关联，实现特征的渐进式融合与增强。实验结果显示，DPCMNet在USOD（Unicorn Small Object Dataset）和VEDAI（VEhicle Detection in Aerial Imagery）遥感小目标数据集上均提升了检测性能，该网络的平均精度（AP）较基准模型YOLOv11n分别提高了2.9和3.1个百分点。

关键词: 遥感图像, 小目标检测, 上下文信息增强, 特征交互, 多尺度融合

CLC Number:

TP751.1

Hua YAO, Gaoming YANG, Xuelian LI, Kaixuan LU. Remote sensing small object detection with dynamic perception and cross-modulation[J]. Journal of Computer Applications, 2026, 46(9): 2968-2976.

姚华, 杨高明, 李雪莲, 陆凯旋. 基于动态感知和交叉调制的遥感小目标检测[J]. 《计算机应用》唯一官方网站, 2026, 46(9): 2968-2976.

Figures/Tables 19

References 23

[1]	袁翔，程塨，李戈，等. 遥感影像小目标检测研究进展［J］. 中国图象图形学报， 2023， 28（6）： 1662-1684.
	Yuan Xiang， Cheng Gong， Li Ge， et al. Progress in small object detection for remote sensing images ［J］. Journal of Image and Graphics， 2023， 28（6）： 1662-1684.
[2]	Su J， Liao J， Gu D， et al. Object detection in aerial images using a multiscale keypoint detection network ［J］. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing， 2021， 14： 1389-1398.
[3]	Zhu W， Wang Z， Wang X， et al. A dual self-attention mechanism for vehicle re-identification ［J］. Pattern Recognition， 2023， 137： No.109258.
[4]	Wang B， Sui H， Ma G， et al. GMODet： a real-time detector for ground-moving objects in optical remote sensing images with regional awareness and semantic-spatial progressive interaction ［J］. IEEE Transactions on Geoscience and Remote Sensing， 2025， 63： No.5605623.
[5]	余俊宇，刘孙俊，许桃. 融合注意力机制的YOLOv7遥感小目标检测算法研究［J］. 计算机工程与应用， 2023， 59（20）： 167-175.
	Yu Junyu， Liu Sunjun， Xu Tao. Research on YOLOv7 remote sensing small target detection algorithm integrating attention mechanism［J］. Computer Engineering and Applications， 2023， 59（20）： 167-175.
[6]	刘赏，周煜炜，代娆，等. 融合注意力和上下文信息的遥感图像小目标检测算法［J］. 计算机应用， 2025， 45（1）： 292-300.
	Liu Shang， Zhou Yuwei， Dai Rao， et al. Small target detection algorithm in remote sensing images integrating attention and contextual information ［J］. Journal of Computer Applications， 2025， 45（1）： 292-300.
[7]	Ren S， He K， Girshick R， et al. Faster R-CNN： towards real-time object detection with region proposal networks ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2017， 39（6）： 1137-1149.
[8]	Sun P， Zhang R， Jiang Y， et al. Sparse R-CNN： end-to-end object detection with learnable proposals ［C］// CVPR 2021. Piscataway： IEEE， 2021： 14449-14458.
[9]	Wang C Y， Bochkovskiy A， Mark Liao H Y. YOLOv7： trainable bag-of-freebies sets new state-of-the-art for real-time object detectors［C］// CVPR 2023. Piscataway： IEEE， 2023： 7464-7475.
[10]	Zhao Y， Lv W， Xu S， et al. DETRs beat YOLOs on real-time object detection ［C］// CVPR 2024. Piscataway： IEEE， 2024： 16965-16974.
[11]	Tan M， Pang R， Le Q V. EfficientDet： scalable and efficient object detection ［C］// CVPR 2020. Piscataway： IEEE， 2020： 10778-10787.
[12]	Deng C， Wang M， Liu L， et al. Extended feature pyramid network for small object detection ［J］. IEEE Transactions on Multimedia， 2022， 24： 1968-1979.
[13]	Yang G， Lei J， Tian H， et al. Asymptotic feature pyramid network for labeling pixels and regions ［J］. IEEE Transactions on Circuits and Systems for Video Technology， 2024， 34（9）： 7820-7829.
[14]	Gao T， Xia S， Liu M， et al. MSNet： multi-scale network for object detection in remote sensing images ［J］. Pattern Recognition， 2025， 158： No.110983.
[15]	Zhang S F， Chi C， Yao Y Q， et al. Bridging the gap between anchor-based and anchor-free detection via adaptive training sample selection ［C］// CVPR 2020. Piscataway： IEEE， 2020： 9756-9765.
[16]	Feng C， Zhong Y， Gao Y， et al. TOOD： task-aligned one-stage object detection ［C］// CVPR 2021. Piscataway： IEEE， 2021： 3490-3499.
[17]	Liu S， Li F， Zhang H， et al. DAB-DETR： dynamic anchor boxes are better queries for DETR ［PP/OL］. V4. arXiv （2022-03-30）［2025-08-10］..
[18]	Zhang Y， Ye M， Zhu G， et al. FFCA-YOLO for small object detection in remote sensing images ［J］. IEEE Transactions on Geoscience and Remote Sensing， 2024， 62： No.5611215.
[19]	Xu Y， He W， Zhang G， et al. From weak textures to dense arrangements： leveraging prior knowledge for small-object detection in remote sensing images ［J］. IEEE Transactions on Geoscience and Remote Sensing， 2025， 63： No.3000915.
[20]	Hu J， Shen L， Sun G. Squeeze-and-excitation networks ［C］// CVPR 2018. Piscataway： IEEE， 2018： 7132-7141.
[21]	Wang Q， Wu B， Zhu P， et al. ECA-Net： efficient channel attention for deep convolutional neural networks ［C］// CVPR 2020. Piscataway： IEEE， 2020： 11531-11539.
[22]	Woo S， Park J， Lee J Y， et al. CBAM： convolutional block attention module ［C］// ECCV 2018， LNCS 11211. Cham： Springer， 2018： 3-19.
[23]	Varghese R， Sambath M. YOLOv8： a novel object detection algorithm with enhanced performance and robustness ［C］// ADICS 2024. Piscataway： IEEE， 2024： 1-6.

模型	AP/%	AP₅₀/%	AP₇₅/%	AP_vt/%	AP_t/%	AP_s/%	参数量/10⁶	浮点运算量/GFLOPs
Faster R-CNN	31.2	81.4	12.3	9.6	30.7	38.3	43.1	223.6
ATSS	31.3	84.9	11.3	9.7	30.8	38.7	32.0	205.2
TOOD	31.6	83.6	12.3	12.2	31.4	37.3	31.9	201.1
DAB-DETR	28.6	80.8	10.4	8.2	28.5	35.3	47.2	163.5
YOLOv7	25.6	79.4	6.8	7.2	25.5	28.0	6.0	13.0
YOLOv8s	29.3	82.1	10.3	11.3	28.7	34.4	11.1	28.8
YOLOv8n	29.8	82.2	11.2	9.6	28.9	36.0	3.1	8.9
YOLOv11s	32.4	87.0	13.7	13.6	31.3	36.9	9.4	21.3
YOLOv11n	31.6	83.2	11.9	12.4	30.8	37.2	2.6	6.6
FFCA-YOLO	33.8	88.8	14.2	15.5	34.1	37.0	7.1	51.7
DPCMNet	34.5	89.4	14.9	15.5	33.3	38.3	1.2	9.9

模型	AP/%	AP₅₀/%	AP₇₅/%	AP_vt/%	AP_t/%	AP_s/%	参数量/10⁶	浮点运算量/GFLOPs
Faster R-CNN	31.2	81.4	12.3	9.6	30.7	38.3	43.1	223.6
ATSS	31.3	84.9	11.3	9.7	30.8	38.7	32.0	205.2
TOOD	31.6	83.6	12.3	12.2	31.4	37.3	31.9	201.1
DAB-DETR	28.6	80.8	10.4	8.2	28.5	35.3	47.2	163.5
YOLOv7	25.6	79.4	6.8	7.2	25.5	28.0	6.0	13.0
YOLOv8s	29.3	82.1	10.3	11.3	28.7	34.4	11.1	28.8
YOLOv8n	29.8	82.2	11.2	9.6	28.9	36.0	3.1	8.9
YOLOv11s	32.4	87.0	13.7	13.6	31.3	36.9	9.4	21.3
YOLOv11n	31.6	83.2	11.9	12.4	30.8	37.2	2.6	6.6
FFCA-YOLO	33.8	88.8	14.2	15.5	34.1	37.0	7.1	51.7
DPCMNet	34.5	89.4	14.9	15.5	33.3	38.3	1.2	9.9

骨干网络	AP	AP₅₀	AP₇₅	AP_vt	AP_t	AP_s
CSPDarkNet53	31.7	86.1	12.8	13.4	30.5	36.6
本文骨干网络	34.5	89.4	14.9	15.5	33.3	38.3

骨干网络	AP	AP₅₀	AP₇₅	AP_vt	AP_t	AP_s
CSPDarkNet53	31.7	86.1	12.8	13.4	30.5	36.6
本文骨干网络	34.5	89.4	14.9	15.5	33.3	38.3

特征融合策略	AP	AP₅₀	AP₇₅	AP_s
BiFPN	34.0	88.5	13.9	36.7
AFPN	29.9	83.9	10.4	32.6
DMFP	34.5	89.4	14.9	38.3

Remote sensing small object detection with dynamic perception and cross-modulation

基于动态感知和交叉调制的遥感小目标检测

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 19

References 23

Related Articles 15

Recommended Articles

Metrics

模型	AP	AP₅₀	AP₇₅	AP_t	AP_s
Faster R-CNN	41.9	70.9	43.0	24.5	42.3
ATSS	40.8	67.9	46.1	35.3	41.1
TOOD	39.8	63.3	43.6	24.6	40.8
DAB-DETR	35.4	67.0	34.0	22.5	35.1
YOLOv7	32.2	55.2	35.3	30.0	33.8
YOLOv8n	40.8	64.6	43.8	30.7	41.8
YOLOv11n	39.6	63.7	42.7	21.6	40.5
FFCA-YOLO	40.0	68.3	44.3	27.0	41.2
DPCMNet	42.7	68.4	48.3	31.4	43.8

模型	浮点运算量/GFLOPs	帧率/（frame·s^-1）	推理时间/ms
YOLOv8s	28.8	164.6	6.0
YOLOv11s	21.3	166.8	5.9
FFCA-YOLO	51.7	61.7	15.0
DPCMNet	9.9	220.4	4.5

g	AP/%	AP₅₀/%	AP₇₅/%	AP_vt/%	AP_t/%	AP_s/%
1	31.0	84.6	11.8	12.7	30.0	35.6
2	34.0	88.9	15.0	16.3	32.8	37.0
3	34.5	89.4	14.9	15.5	33.3	38.3
4	34.2	89.4	14.7	14.5	33.1	37.0
5	33.4	88.0	14.2	16.9	32.1	36.0

m	AP/%	AP₅₀/%	AP₇₅/%	AP_vt/%	AP_t/%	AP_s/%
1	33.6	87.8	14.4	15.3	32.1	36.4
2	34.5	89.4	14.9	15.5	33.3	38.3
3	31.2	85.0	12.0	12.6	30.4	37.5

n	AP/%	AP₅₀/%	AP₇₅/%	AP_vt/%	AP_t/%	AP_s/%
1	33.9	88.5	15.0	15.6	32.7	36.0
2	34.5	89.4	14.9	15.5	33.3	38.3
3	30.9	84.4	12.2	14.3	29.9	35.3

HFA	DMF	CSG	AP	AP₅₀	AP₇₅
×	×	×	30.3	83.5	12.0
√	×	×	31.5	86.2	12.7
×	√	×	32.6	87.5	12.6
√	√	×	33.6	87.9	14.3
√	√	√	34.5	89.4	14.9

层序号	DPCMNet		Baseline
层序号	模块	参数量	模块	参数量
合计	—	704 054	—	1 398 312
0	CBS	464	Conv	464
1	CBS	4 672	Conv	4 672
2	HFA	75 008	C3k2	6 728
3	CBS	18 496	Conv	36 992
4	CSG	77 670	C3k2	26 448
5	CBS	147 712	Conv	147 712
6	HFA	305 664	C3k2	117 376
7	SPPF	74 368	Conv	295 424
8	—	—	C3k2	335 872
9	—	—	SPPF	164 480
10	—	—	C2PSA	262 144

层序号	模块	参数量
合计	—	547 907
8	Concat（6，10）	0
9	DMF	115 395
10	CBS	18 496
11	Concat（4，9，10）	0
12	DMF	35 427
13	CBS	36 992
14	Concat（7，9，13）	0
15	DMF	123 587
16	Detection Head	218 010

[1]	He SUN, Guanghui YAN, Xiaohong JIA. Superpixel segmentation algorithm based on inverted feature pyramid network [J]. Journal of Computer Applications, 2026, 46(9): 2959-2967.
[2]	Wenling LIU, Yong LI, Jiahui LI, Houkang ZHANG. Unsupervised dehazing method for remote sensing images with non-uniform haze [J]. Journal of Computer Applications, 2026, 46(7): 2297-2306.
[3]	Minqi WU, Yuanhua YANG, Hang LI, Yaqin HU, Zhihao TANG, Teng MEI. Lightweight underwater small object detection based on graph Transformer and RT-DETR [J]. Journal of Computer Applications, 2026, 46(5): 1586-1595.
[4]	Jiaxin DUAN, Jing HU, Wu WEN, Zhenxia YU, Yongjun ZHANG. Pansharpening based on two-stage collaborative optimization [J]. Journal of Computer Applications, 2026, 46(5): 1568-1577.
[5]	Hongrui ZHANG, Weiming FENG, Luxia YANG, Yongjie MA. CSAF-YOLO： improved YOLO11 algorithm for underwater small object detection [J]. Journal of Computer Applications, 2026, 46(5): 1578-1585.
[6]	Hanqing LIU, Guoming SANG, Yijia ZHANG. Remote sensing image captioning model combining dense multi-scale feature fusion and feature knowledge-enhanced Transformer [J]. Journal of Computer Applications, 2026, 46(3): 741-749.
[7]	Yanqun LU, Yiyi ZHAO. Customer churn prediction model integrating hierarchical graph neural network and specific feature learning [J]. Journal of Computer Applications, 2025, 45(9): 3057-3066.
[8]	Yingtao CHEN, Kangkang FANG, Jin’ao ZHANG, Haoran LIANG, Huanbin GUO, Zhaowen QIU. Segmentation network of coronary artery structure from CT angiography images based on multi-scale spatial features [J]. Journal of Computer Applications, 2025, 45(6): 2007-2015.
[9]	Yang ZHOU, Hui LI. Remote sensing image building extraction network based on dual promotion of semantic and detailed features [J]. Journal of Computer Applications, 2025, 45(4): 1310-1316.
[10]	Zhongwei ZHANG, Jun WANG, Shudong LIU, Zhiheng WANG. Object detection in remote sensing image based on multi-scale feature fusion and weighted boxes fusion [J]. Journal of Computer Applications, 2025, 45(2): 633-639.
[11]	Songsen YU, Zhifan LIN, Guopeng XUE, Jianyu XU. Lightweight large-format tile defect detection algorithm based on improved YOLOv8 [J]. Journal of Computer Applications, 2025, 45(2): 647-654.
[12]	Feng YUE, Yang PENG, Zhaopin SU, Guofu ZHANG, Chensi LIAN, Bo YANG, Zhen FANG. Open-set source cell-phone identification method based on feature interaction and representation enhancement [J]. Journal of Computer Applications, 2025, 45(12): 3813-3819.
[13]	Yonghong FAN, Heming HUANG. CnnPRL： progressive representation learning method for speech emotion recognition [J]. Journal of Computer Applications, 2025, 45(12): 3804-3812.
[14]	Shang LIU, Yuwei ZHOU, Rao DAI, Linfang DONG, Meng LIU. Small target detection algorithm in remote sensing images integrating attention and contextual information [J]. Journal of Computer Applications, 2025, 45(1): 292-300.
[15]	Chenqian LI, Jun LIU. Ultrasound carotid plaque segmentation method based on semi-supervision and multi-scale cascaded attention [J]. Journal of Computer Applications, 2024, 44(8): 2604-2610.