[1]李北明,金荣璐,徐召飞,等.基于特征蒸馏的改进Ghost-Yolov5红外目标检测算法[J].郑州大学学报(工学版),2022,43(01):20-26.[doi:10.13705/j.issn.1671-6833.2022.01.013]
 LI Beiming,JIN Ronglu,XU Zhaofei,et al.An Improved Ghost-YOLOv5 Infrared Target Detection Algorithm Based on Feature Distillation[J].Journal of Zhengzhou University (Engineering Science),2022,43(01):20-26.[doi:10.13705/j.issn.1671-6833.2022.01.013]
点击复制

基于特征蒸馏的改进Ghost-Yolov5红外目标检测算法()
分享到:

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
43
期数:
2022年01期
页码:
20-26
栏目:
出版日期:
2022-01-09

文章信息/Info

Title:
An Improved Ghost-YOLOv5 Infrared Target Detection Algorithm Based on Feature Distillation
作者:
李北明1金荣璐12徐召飞2刘晴2王水根2
哈尔滨工程大学信息与通信工程学院;烟台艾睿光电科技有限公司;

Author(s):
LI Beiming1 JIN Ronglu12 XU Zhaofei2 LIU Qing2 WANG Shuigen2
1.College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China; 
2.Yantai IRay Technology Co., Ltd., Yantai 264000, China
关键词:
Keywords:
infrared target detection data enhancement model pruning feature distillation Hisi plat-form
分类号:
TP391.9
DOI:
10.13705/j.issn.1671-6833.2022.01.013
文献标志码:
A
摘要:
红外目标检测算法一直是重要的研究领域之一,广泛应用于安防、军事等领域。当前大部分目标检测算法存在计算复杂度高、准确率低、适应性差等问题,难以满足实际应用需求,以及部署于移动嵌入式平台设备。针对上述问题,本文提出了一种基于特征蒸馏的改进Ghost-YOLOv5红外目标检测算法,该算法首先在YOLOv5模型结构的基础上利用Ghost模块做模型剪枝,然后继续对模型进行通道剪枝;其次使用Mosaic和Copy-paste两种数据增强方法和特征蒸馏方法提高压缩后模型的检测精度。进一步地,考虑到安防红外目标检测领域尚无开源的数据集,本文构建了一个包含多种场景下人、机动车和非机动车目标的数据集(即将开源)。在上述数据集上测试实验结果表明:本文提出的算法利用Ghost模块和通道裁剪可以把原始的YOLOv5模型大小减少到1.9M,相比于原模型大小减少了93%,模型复杂度(GFLOPS)降低了86.1%,并通过知识蒸馏和数据增强的方法,使得裁剪后的模型在红外数据集上的精度提升了6.9,总体mAP值达到了90.4,比YOLOv5模型高0.3。在型号为Hi3519AV100的海思平台上实测,模型的检测速度能达到25帧,平均检测精度能达到90.2,比YOLOv5s高0.4,比YOLOv3高4.8,比YOLOv3-tiny高14.4,比YOLOv4-tiny高9.6。
Abstract:
Infrared target detection algorithms suffered from problems such as poor adaptability and high computational complexity. An improved Ghost-YOLOv5 infrared target detection algorithm was proposed based on feature distillation to solve the above problems. Firstly, GhostNet block was used for backbone pruning. Se-condly, two effective data enhancement methods including Mosaic and Copy-paste were used, together with feature distillation to improve the accuracy in object detection. Furthermore, an infrared image dataset that contained a variety of scenarios with pedestrians, motor vehicles, and non motorized targets was constructed. The test experimental results on the above dataset showed that the model parameters obtained by the algorithm proposed in this paper using GhostNet module were only 1.9M, and the accuracy of the small model on the infrared dataset were improved by 6.6% through feature distillation and data enhancement. And the overall mAP value reached 90.1%. The detection speed of the model could reach 25 frames per second and the average detection accuracy could reach 90.2% when measured empirically on Hisi, all achieving higher detection accuracy compared to a variety of common models portable to this platform.

参考文献/References:

[1] 张敏,韩芳,康键,等.红外热成像技术在民用领域 的应用[J].红外,2019,40( 6) : 37-45. 

[2] 陈华,周晓巍.红外热成像技术在通信电源电路板 故障检 测 中 的 应 用[J]. 通 信 电 源 技 术,2021,38 ( 1) : 114-116. 
[3] 王洪琳,郑睿.激光雷达在轨道列车自动驾驶系统 中的应用研究[J].江苏科技信息,2021,38 ( 31) : 42-44,55. 
[4] 蔡军,黄袁园,李鹏泽,等.基于视觉对比度机制的 红外弱小目标检测算法[J].系统工程与电子技术, 2019,41( 11) : 2416-2423. 
[5] KIM S,YANG Y,LEE J,et al. Small target detection utilizing robust methods of the human visual system for IRST[J].Journal of infrared,millimeter and terahertz waves,2009,30( 9) : 994-1011.
 [6] SHAO X P,FAN H,LU G X,et al. An improved infrared dim and small target detection algorithm based on the contrast mechanism of human visual system[J]. Infrared physics & technology,2012,55( 5) : 403-408. 
[7] CHEN C L P,LI H,WEI Y T,et al.A local contrast method for small infrared target detection[J]. IEEE transactions on geoscience and remote sensing,2014, 52( 1) : 574-581. 
[8] 张震,李浩方,李孟洲,等.改进 YOLOv3 算法与人体 信息数据融合的视频监控检测方法[J].郑州大学 学报( 工学版) ,2021,42( 1) : 28-34. 
[9] 崔娟.红外偏振成像技术在船舶目标探测中的应用 [J].舰船科学技术,2020,42( 18) : 82-84. 
[10] DU L,GAO C Q,FENG Q,et al.Small UAV detection in videos from a single moving camera[C]/ /Communications in Computer and Information Science. Cham: Springer,2017: 187-197. 
[11] CHEN P H,LIN C J,SCHÖLKOPF B. A tutorial on ν-support vector machines [J]. Applied stochastic models in business and industry,2005,21 ( 2) : 111 -136.
 [12] 汪慎文,张佳星,褚晓凯,等.两阶段搜索的多模态 多目标差分进化算法[J].郑州大学学报( 工学版) , 2021,42( 1) : 9-14,110. 
[13] 周贵华,许丽娟,周伟昌.基于深度学习的红外序列 图像小目标检测方法研究[J].激光杂志,2020,41 ( 12) : 61-64.
 [14] HE K,ZHANG X,REN S,et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE transactions on pattern analysis and machine intelligence,2015,37( 9) : 1904-1916. 
[15] LIN T Y,DOLLAR P,GIRSHICK R,et al.Feature pyramid networks for object detection[C]/ /2017 IEEE Conference on Computer Vision and Pattern Recognition ( CVPR) .Piscataway: IEEE,2017: 2117-2125. 
[16] LIU S,QI L,QIN H F,et al.Path aggregation network for instance segmentation[C]/ /2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE,2018: 8759-8768. 
[17] 陈寂驰,魏国华,郭聪隆,等.一种基于红外图像序 列的深度学习三维重建仿真方法初探[J].空天防 御,2020,3( 4) : 21-29. 
[18] HAN K,WANG Y H,TIAN Q,et al. GhostNet: more features from cheap operations[C]/ /2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition ( CVPR) . Piscataway: IEEE,2020: 1580 -1589. 
[19] YIM J,JOO D,BAE J,et al.A gift from knowledge distillation: fast optimization,network minimization and transfer learning[C]/ /2017 IEEE Conference on Computer Vision and Pattern Recognition ( CVPR) . Piscataway: IEEE,2017: 4133-4141. 
[20] WANG C Y,BOCHKOVSKIY A,LIAO H Y M.ScaledYOLOv4: scaling cross stage partial network[EB/OL]. ( 2020-11- 16) [2021 - 05 - 06]. https: / /arxiv. org / abs/2011. 08036. 
[21] RUDER S.An overview of gradient descent optimization algorithms[EB/OL]. ( 2016-09-15) [2021-05-06]. https: / /arxiv.org /abs/1609. 04747.

更新日期/Last Update: 2022-01-09