改进YOLOv5的轻量化红外交通目标检测

doi:10.3778/j.issn.1002-8331.2301-0190

摘要/Abstract

摘要： 针对如何提高在复杂场景下红外目标检测性能以及如何平衡算法的轻量化与精确度等问题，提出了一种基于YOLOv5的轻量化红外目标检测算法。该算法设计一种融合并行卷积的Ghost特征提取模块，利用结构重参化的思想，在推理时将模型简化，降低模型的复杂度且加快推理速度；同时对网络进行剪枝，增加了小目标检测层，提高了模型对复杂场景的检测性能；在特征融合模块设计了一种混洗Ghost模块，将Ghost特征和标准卷积特征进行混洗，尽可能减少Ghost特征对网络性能的负面影响；设计了一个解耦检测头模块，将分类与检测任务进行分解，提高了网络在复杂环境下的定位检测能力。实验结果表明，与YOLOv5s相比，提出的算法mAP提高了4.7个百分点，参数量降低了42.7%，推理延时减少了8.5%，能够在检测性能和轻量化上达到理想平衡。

关键词: 红外目标检测, YOLOv5, 幽灵网络, 结构重参化

Abstract: Considering the problems of how to improve the performance of IR object detection in complex scenes and how to balance the lightness and accuracy of the algorithm, this paper proposes a lightweight IR target detection algorithm based on YOLOv5. This paper designs a Ghost feature extraction module that incorporates parallel convolution to reduce the complexity of the model and speed up the inference by using the idea of structural re-referencing to simplify the model at the time of inference. The network is pruned to add small target detection layer and improve the detection performance in complex scenes. Secondly, a Shuffle Ghost module is designed in the feature fusion module to shuffle ghost features and standard convolutional features to minimize the negative impact of ghost features on network performance. Finally, a decoupled detection head module is designed to decompose the classification and detection tasks to improve the localization and detection capability of the network in complex environments. The experimental results show that proposed algorithm has a 4.7 percentage points improvement in mAP, a 42.7% reduction in the parameters, and an 8.5% reduction in inference latency compared to YOLOv5s, and is able to achieve an ideal balance of detection performance and lightweight deployment.

Key words: IR object detection, YOLOv5, ghost net, structural re-referencing

邓楷文, 葛晨阳. 改进YOLOv5的轻量化红外交通目标检测[J]. 计算机工程与应用, 2023, 59(12): 184-192.

DENG Kaiwen, GE Chenyang. Research on Lightweight of Improved YOLOv5 Infrared Traffic Detection Network[J]. Computer Engineering and Applications, 2023, 59(12): 184-192.

参考文献

[1] 钱伍，王国中，李国平.改进YOLOv5的交通灯实时检测鲁棒算法[J].计算机科学与探索，2022，16（1）：231-241.
QIAN W，WANG G Z，LI G P.Improved YOLOv5 traffic light real-time detection robust algorithm[J].Journal of Frontiers of Computer Science and Technology，2022，16（1）：231-241.
[2] 宋艳艳，谭励，马子豪，等.改进YOLOV3算法的视频目标检测[J].计算机科学与探索，2021，15（1）：163-172.
SONG Y Y，TAN L，MA Z H，et al.Video target detection based on improved YOLOV3 algorithm[J].Journal of Frontiers of Computer Science and Technology，2021，15（1）：163-172.
[3] 卢湖川，李佩霞，王栋.目标跟踪算法综述[J].模式识别与人工智能，2018，31（1）：61-76.
LU H C，LI P X，WANG D.Visual object tracking：a survey[J].Pattern Recognition and Artificial Intelligence，2018，31（1）：61-76.
[4] 李俊宏，张萍，王晓玮.红外弱小目标检测算法综述[J].中国图象图形学报，2020，25（9）：1739-1753.
LI J H，ZHANG P，WANG X W.Infrared small-target detection algorithms：a survey[J].Journal of Image and Graphics，2020，25（9）：1739-1753.
[5] SEYED M F，REZA M M，MAHDI N.Flying small target detection in IR images based on adaptive toggle operator[J].IET Computer Vision，2018，12（4）：527-534.
[6] 吴天爱，黄树彩，苑智玮.联合NSCT和SVD方法的红外弱小目标复杂背景抑制[J].红外技术，2016，38（9）：758-764.
WU T A，HUANG S C，YUAN Z W.NSCT combined with SVD for infrared dim target complex background suppression[J].Infrared Technology，2016，38（9）：758-764.
[7] WANG X，PENG Z，ZHANG P，et al.Infrared small target detection via nonnegativity-constrained variational mode decomposition[J].IEEE Geoscience & Remote Sensing Letters，2017，14（10）：1700-1704.
[8] WEI Y，YOU X，LI H.Multiscale patch-based contrast measure for small infrared target detection[J].Pattern Recognition，2016，58：216-226.
[9] DAI Y M，WU Y，SONG Y.Infrared small target and background separation via column-wise weighted robust principal component analysis[J].Infrared Physics & Technology，2016，77：421-430.
[10] 邵延华，张铎，楚红雨，等.基于深度学习的YOLO目标检测综述[J].电子与信息学报，2022，44（10）：12-18.
SHAO Y H，ZHANG D，CHU H Y，et al.A review of YOLO object detection based on deep learning[J].Journal of Electronics & Information Technology，2022，44（10）：12-18.
[11] 王恒慧，曹东，赵杨.基于深度学习的红外弱小目标检测算法研究综述[J].激光与红外，2022，52（9）：1274-1279.
WANG H H，CAO D，ZHAO Y，Survey of infrared dim small target detection algorithm based on deep learning[J].Laser & Infrared，2022，52（9）：1274-1279.
[12] 刘学，李范鸣，刘士建.改进的SSD红外图像行人检测算法[J].电光与控制，2020，27（1）：42-46.
LIU X，LI F M，LIU S J.An infrared image pedestrian detection algorithm based on improved SSD algorithm[J].Electronics Optics & Control，2020，27（1）：42-46.
[13] PEI D，JING M，LIU H，et al.A fast RetinaNet fusion framework for multi-spectral pedestrian detection[J].Infrared Physics & Technology，2020，105：103178.
[14] IVASIC-KOS M，KRISTO M，POBAR M.Person detection in thermal videos using YOLO[C]//Proceedings of the Intelligent Systems and Applications，2020.
[15] SHI J，ZHANG G，YUAN J，et al.Improved YOLOv3 infrared image pedestrian detection algorithm[C]//International Conference of Pioneering Computer Scientists，Engineers and Educators，2020.
[16] HOWARD A G，ZHU M，CHEN B，et al.Mobilenets：efficient convolutional neural networks for mobile vision applications[J].arXiv：1704.04861，2017.
[17] SANDLER M，HOWARD A，ZHU M，et al.Mobilenetv2：inverted residuals and linear bottlenecks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018.
[18] ZHANG X，ZHOU X，LIN M，et al.Shufflenet：an extremely efficient convolutional neural network for mobile devices[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018.
[19] MA N，ZHANG X，ZHENG H T，et al.Shufflenet v2：practical guidelines for efficient CNN architecture design[C]//Proceedings of the European Conference on Computer Vision，2018.
[20] HAN K，WANG Y，TIAN Q，et al.Ghostnet：more features from cheap operations[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020.
[21] DING X，ZHANG X，MA N，et al.RepVGG：making VGG-style convnets great again[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2021.
[22] 赵凤，李永恒，李晶.基于改进YOLOv4-tiny的轻量化室内人员目标检测算法[J].电子与信息学报，2022，44（11）：3815-3824.
ZHAO F，LI Y H，LI J.Lightweight indoor personnel detection algorithm based on improved YOLOv4-tiny[J].Journal of Electronics & Information Technology，2022，44（11）：3815-3824.
[23] MUDASSAR B A，MUKHOPADHYAY S.Rethinking convolutional feature extraction for small object detection[C]//Proceedings of the BMVC，2019.