结合特征融合增强和细节特征的低照度小目标检测方法

江泽涛; 程留明; 杨建琛

doi:10.12263/DZXB.20241155

您当前的位置：

首页 >

文章列表页 >

结合特征融合增强和细节特征的低照度小目标检测方法

学术论文 | 更新时间：2025-12-10

- 结合特征融合增强和细节特征的低照度小目标检测方法
- Low-Light Small Target Detection Method Combining Feature Fusion Enhancement and Detail Features
- 电子学报 2025年53卷第7期页码：2229-2240
- 作者机构：
  
  桂林电子科技大学广西图像图形处理与智能处理重点实验室，广西桂林 541004
- 作者简介：
  
  [ "江泽涛男，1961年3月出生于江西省九江市.现为桂林电子科技大学教授、博士生导师.主要研究方向为图像处理、计算机视觉、人工智能.中国知网高被引学者TOP1%. E-mail: 394503704@qq.com" ]
  [ "程留明男，1999年1月出生于河南省周口市.现为桂林电子科技大学硕士研究生.主要研究方向为计算机视觉、低照度小目标检测.E-mail: 22032201007@mails.guet.edu.cn" ]
  [ "杨建琛男，1998年5月出生于河北省张家口市.现为桂林电子科技大学硕士研究生.主要研究方向为计算机视觉、低照度图像增强.E-mail: 22032202038@mails.guet.edu.cn" ]
- 基金信息：
  
  国家自然科学基金(62473105);广西自然学科基金重点项目(2021GXNSFDA196002);广西图像图形智能处理重点实验项目(GIIP2302;GIIP2303;GIIP2304);桂林电子科技大学研究生教育创新计划项目(2024YCXB09;2024YCXS039;2024YCXS035)
- DOI：10.12263/DZXB.20241155
  中图分类号： TP391;
- 收稿：2024-12-24，
  
  修回：2025-04-28，
  
  纸质出版：2025-07-25
- 稿件说明：
移动端阅览
江泽涛, 程留明, 杨建琛. 结合特征融合增强和细节特征的低照度小目标检测方法[J]. 电子学报, 2025, 53(07): 2229-2240.

JIANG Ze-tao, CHENG Liu-ming, YANG Jian-chen. Low-Light Small Target Detection Method Combining Feature Fusion Enhancement and Detail Features[J]. Acta Electronica Sinica, 2025, 53(07): 2229-2240.
江泽涛, 程留明, 杨建琛. 结合特征融合增强和细节特征的低照度小目标检测方法[J]. 电子学报, 2025, 53(07): 2229-2240. DOI：10.12263/DZXB.20241155

JIANG Ze-tao, CHENG Liu-ming, YANG Jian-chen. Low-Light Small Target Detection Method Combining Feature Fusion Enhancement and Detail Features[J]. Acta Electronica Sinica, 2025, 53(07): 2229-2240. DOI：10.12263/DZXB.20241155

摘要

正常照度环境下的图像小目标检测具有挑战性，而低照度环境下的图像因低亮度、低对比度、低信噪比等，信息丢失严重，使得小目标特征信息更加弱化，更加难以获取特征等信息，从而导致低照度环境下小目标检测的研究极少.本文针对这一问题，提出一个低照度小目标检测方法MC-YOLO.该方法包含多尺度特征融合增强（Multi-Scale Feature Fusion Enhancement，MFFE）模块、细节特征提取（Detail Feature Extraction，DFE）模块、Neck模块和Head模块.该方法的第一部分由MFFE模块采用可变形卷积和多尺度特征融合来提取低照度环境下的小目标特征，再通过全局平均池化进行全局特征增强，从而使小目标特征信息更加显著；第二部分由DFE模块在充分利用上下文信息提取小目标特征的同时保留小目标的位置信息，解决低照度小目标物体细节特征信息易丢失的问题；第三部分由Neck模块进行特征提取和多尺度特征融合；最后，在Head模块的大分辨率特征图上引入小目标检测层，实现低照度小目标的检测.实验结果表明，本文方法在低照度小目标检测精度方面具有良好的表现，在自制低照度小目标数据集LLSOD上的mAP为83.2%，比目前先进的目标检测方法YOLOv11高出3.6%.

Abstract

The detection of small targets in normal illumination conditions is challenging. In low-light environments

images suffer from severe information loss due to low brightness

low contrast

and low signal-to-noise ratio

which further weakens the feature information of small targets

making feature extraction more difficult. As a result

research on small target detection in low-light environments is scarce. To address this issue

this paper proposes a low-light small target detection method

MC-YOLO. The MC-YOLO includes four modules: multi-scale feature fusion enhancement (MFFE) module

detail feature extraction (DFE) module

Neck module and Head module. Firstly

the method uses the MFFE module to extract small target features in low-light environment through deformable convolutions and multi-scale feature fusion and to enhance global features through global average pooling making small target feature information more salient. Next

the DFE module fully utilizes contextual information to extract small target features while preserving the positional information of the small targets

which solves the problem of easy loss of detail feature information of small targets in low-light environments. Then

the neck module performs feature extraction and multi-scale feature fusion. Finally

the head module introduces a small target detection layer on the high-resolution feature map to detect small targets in low-light environments. Experimental results show that this method performs well in the accuracy of low-light small target detection

with the mAP of 83.2% on the self-made low-light small target dataset LLSOD

which is 3.6% higher than the current advanced target detection method YOLOv11.

关键词

Keywords

references

KANG S H , PARK J S . Aligned matching: Improving small object detection in SSD [J ] . Sensors , 2023 , 23 ( 5 ): 2589 .

LIU H , DING M , LI S , et al . Small-target detection based on an attention mechanism for apron-monitoring systems [J ] . Applied Sciences , 2023 , 13 ( 9 ): 5231 .

BETTI A , TUCCI M . YOLO-S: A lightweight and accurate YOLO-like network for small target selection in aerial imagery [J ] . Sensors , 2023 , 23 ( 4 ): 1865 .

HU M Z , LI Z Y , YU J , et al . Efficient-lightweight YOLO: Improving small object detection in YOLO for aerial images [J ] . Sensors , 2023 , 23 ( 14 ): 6423 .

QI G Q , ZHANG Y C , WANG K P , et al . Small object detection method based on adaptive spatial parallel convolution and fast multi-scale fusion [J ] . Remote Sensing , 2022 , 14 ( 2 ): 420 .

JI C L , YU T , GAO P , et al . Yolo-tla: An efficient and lightweight small object detection model based on YOLOv5 [J ] . Journal of Real-Time Image Processing , 2024 , 21 ( 4 ): 141.1 - 141.16 .

WANG T Q , QU H Q , LIU C A , et al . LLE-STD: Traffic sign detection method based on low-light image enhancement and small target detection [J ] . Mathematics , 2024 , 12 ( 19 ): 3125 .

DAI T , WANG Q , SHEN Y C , et al . SwinVision: Detecting small objects in low-light environments [J ] . IEEE Access , 2025 , 13 : 42797 - 42812 .

DONG X , WANG G , PANG Y , et al . Fast efficient algorithm for enhancement of low lighting video [C ] // 2011 IEEE International Conference on Multimedia and Expo . Piscataway : IEEE , 2011 : 1 - 6 .

PARTHASARATHY S , SANKARAN P . Fusion based multi scale RETINEX with color restoration for image enhancement [C ] // 2012 International Conference on Computer Communication and Informatics . Piscataway : IEEE , 2012 : 1 - 7 .

ARICI T , DIKBAS S , ALTUNBASAK Y . A histogram modification framework and its application for image contrast enhancement [J ] . IEEE Transactions on Image Processing , 2009 , 18 ( 9 ): 1921 - 1935 .

GUO X J , LI Y , LING H B . LIME: Low-light image enhancement via illumination map estimation [J ] . IEEE Transactions on Image Processing , 2017 , 26 ( 2 ): 982 - 993 .

YANG S Z , DING M X , WU Y M , et al . Implicit neural representation for cooperative low-light image enhancement [C ] // 2023 IEEE/CVF International Conference on Computer Vision (ICCV) . Piscataway : IEEE , 2023 : 12872 - 12881 .

ZHANG H , GOODFELLOW I , METAXAS D , et al . Self-attention generative adversarial networks [EB/OL ] . ( 2018-05-21 )[ 2025-05-05 ] . https://arxiv.org/abs/1805.08318v2 https://arxiv.org/abs/1805.08318v2 .

HOU Q B , ZHOU D Q , FENG J S . Coordinate attention for efficient mobile network design [C ] // 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2021 : 13708 - 13717 .

PARK J , WOO S , LEE J Y , et al . Bam: Bottleneck attention module [C ] // British Machine Vision Conference . Newcastle : BMVC , 2018 : 147 - 160 .

AZAD R , NIGGEMEIER L , HÜTTEMANN M , et al . Beyond self-attention: Deformable large kernel attention for medical image segmentation [C ] // 2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) . Piscataway : IEEE , 2024 : 1276 - 1286 .

WANG C Y , YEH I H , MARK LIAO H Y . YOLOv9: Learning what you want to learn using programmable gradient information [M ] // Computer Vision - ECCV 2024 . Cham : Springer Nature Switzerland , 2024 : 1 - 21 .

REN S Q , HE K M , 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 .

LIN T Y , GOYAL P , GIRSHICK R , et al . Focal loss for dense object detection [C ] // 2017 IEEE International Conference on Computer Vision (ICCV) . Piscataway : IEEE , 2017 : 2999 - 3007 .

TIAN Z , SHEN C H , CHEN H , et al . FCOS: A simple and strong anchor-free object detector [J ] . IEEE Transactions on Pattern Analysis and Machine Intelligence , 2022 , 44 ( 4 ): 1922 - 1933 .

XU S L , WANG X X , LV W Y , et al . PP-YOLOE: An evolved version of YOLO [EB/OL ] . ( 2022-12-12 )[ 2025-05-05 ] . https://arxiv.org/abs/2203.16250v3 https://arxiv.org/abs/2203.16250v3 .

JIA D , YUAN Y H , HE H D , et al . DETRs with hybrid matching [C ] // 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2023 : 19702 - 19712 .

WANG A , CHEN H , LIU L H , et al . YOLOv10: Real-time end-to-end object detection [EB/OL ] . ( 2024-12-30 )[ 2025-05-05 ] . https://arxiv.org/abs/2405.14458v2 https://arxiv.org/abs/2405.14458v2 .

CAI Z W , VASCONCELOS N . Cascade R-CNN: Delving into high quality object detection [C ] // 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2018 : 6154 - 6162 .

FENG C J , ZHONG Y J , GAO Y , et al . TOOD: Task-aligned one-stage object detection [C ] // 2021 IEEE/CVF International Conference on Computer Vision (ICCV) . Piscataway : IEEE , 2021 : 3490 - 3499 .

LIU Y F , HE M , HUI B . ESO-DETR: An improved real-time detection transformer model for enhanced small object detection in UAV imagery [J ] . Drones , 2025 , 9 ( 2 ): 143 .

WU M J , YUN L J , WANG Y B , et al . Detection algorithm for dense small objects in high altitude image [J ] . Digital Signal Processing , 2024 , 146 : 104390 .

WANG Z , SU Y T , KANG F , et al . PC-YOLO11s: A lightweight and effective feature extraction method for small target image detection [J ] . Sensors , 2025 , 25 ( 2 ): 348 .

HOU W , WU H M , WU D , et al . Small object detection method for UAV remote sensing images based on αS-YOLO [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2025 , 18 : 8984 - 8994 .

GUO Z , BI G L , LV H Y , et al . No-extra components density map cropping guided object detection in aerial images [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2024 , 62 : 5644813 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于EIMYOLO的高分遥感图像目标检测

面向时序异常检测的可变视距多向扫描方法

基于实时语义链表构建系统的改善定位研究