[1]张 睿,李允臣,王家宝*,等.基于深度学习的红外目标检测综述[J].计算机技术与发展,2023,33(11):1-8.[doi:10. 3969 / j. issn. 1673-629X. 2023. 11. 001]
　ZHANG Rui,LI Yun-chen,WANG Jia-bao*,et al.Survey on Infrared Object Detection Based on Deep Learning[J].,2023,33(11):1-8.[doi:10. 3969 / j. issn. 1673-629X. 2023. 11. 001]

点击复制

基于深度学习的红外目标检测综述()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年11期

页码:

1-8

栏目:

综述

出版日期:

2023-11-10

文章信息/Info

Title:

Survey on Infrared Object Detection Based on Deep Learning

文章编号:

1673-629X(2023)11-0001-08

作者:

张 睿; 李允臣; 王家宝* ; 李 阳; 苗 壮

陆军工程大学 指挥控制工程学院,江苏 南京 210007

Author(s):

ZHANG Rui; LI Yun-chen; WANG Jia-bao* ; LI Yang; MIAO Zhuang

College of Command and Control Engineering,Army Engineering University of PLA,Nanjing 210007,China

关键词:

深度学习; 红外目标检测; 迁移学习; 注意力机制; 特征融合; 图像融合

Keywords:

deep learning; infrared object detection; transfer learning; attention mechanism; feature fusion; image fusion

分类号:

TP391. 4

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 11. 001

摘要:

红外图像由于分辨率低、纹理细节不足,且缺乏颜色信息,导致目标成像模糊,检测难度大。 基于深度学习的红外目标检测技术,通过运用神经网络自动提取复杂的目标特征,大大提高

了检测精度和检测效率,在自动驾驶、安防监控、军事侦察等领域得到了非常广泛的应用。 该文对红外目标检测面临的困难和挑战进行了详细分析,并从数据增强、迁移学习、视觉注意力机制、多尺度特征融合、多模态图像融合和轻量化改进等六个方面,对基于深度学习的红外目标检测研究改进方向进行了系统阐述。 针对红外目标检测数据集缺乏的问题,梳理汇总了 11 个红外目标检测数据集。 同时,结合当前发展现状,对红外目标检测的未来发展方向进行了展望,可为其他研究者提供参考借鉴。

Abstract:

Infrared images have low resolution, insufficient texture details, and lack of color information, resulting in blurred objectimaging and difficult detection. Deep learning-based infrared object detection technology uses complex neural networks to automaticallyextract object features, has greatly improved detection accuracy and detection efficiency,?

and has been widely used in the fields ofautonomous driving,security surveillance,and military reconnaissance. We present a detailed analysis of the difficulties and challenges　facing infrared target detection,and systematically describe the improvement directions of deep learning-based infrared object detectionresearch in six aspects,including data enhancement,transfer learning,visual attention mechanism,multiscale feature fusion,multimodalimage fusion and lightness improvement. To address the problem of lack of infrared object detection datasets,we compose and summarize11 infrared object detection datasets. At the same time,the future development direction of infrared object detection is prospected with thecurrent development status,which can provide reference for other researchers.

常用功能