基于机器学习的MRTD客观测试方法研究

doi:10.12305/j.issn.1001-506X.2024.10.03

系统工程与电子技术 ›› 2024, Vol. 46 ›› Issue (10): 3265-3270.doi: 10.12305/j.issn.1001-506X.2024.10.03

• 电子技术 • 上一篇

基于机器学习的MRTD客观测试方法研究

季然¹^,², 肖茂森¹^,*, 李硕¹^,², 刘宇¹^,³, 罗湛仪¹^,³, 程嘉维¹^,³

1. 中国科学院西安光学精密机械研究所, 陕西西安 710119
2. 中国科学院大学光电学院, 北京 100049
3. 陕西师范大学物理学与信息技术学院, 陕西西安 710119

收稿日期:2023-08-18 出版日期:2024-09-25 发布日期:2024-10-22
通讯作者: 肖茂森
作者简介:季然(1998—), 男, 硕士研究生, 主要研究方向为计算机视觉、红外图像处理、遥感图像处理
肖茂森(1979—), 男, 高级工程师, 硕士, 主要研究方向为空间高精度方位传递对准技术、图像处理
李硕(1994—), 男, 硕士研究生, 主要研究方向为偏振多光谱图像处理、图像超分辨
刘宇(2000—), 女, 硕士研究生, 主要研究方向为信号与图像处理、电子稳像
罗湛仪(1999—), 男, 硕士研究生, 主要研究方向为基于FPGA的图像处理技术
程嘉维(1999—), 男, 硕士研究生, 主要研究方向为红外光学设计、星敏感器

Research on MRTD objective testing method based on machine learning

Ran JI¹^,², Maosen XIAO¹^,*, Shuo LI¹^,², Yu LIU¹^,³, Zhanyi LUO¹^,³, Jiawei CHENG¹^,³

1. Xi'an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Xi'an 710119, China
2. School of Optoelectronics, University of Chinese Academy of Sciences, Beijing 100049, China
3. School of Physics and Information Technology, Shaanxi Normal University, Xi'an 710119, China

Received:2023-08-18 Online:2024-09-25 Published:2024-10-22
Contact: Maosen XIAO

摘要/Abstract

摘要：

红外成像技术的加速发展对红外成像系统测试和评估的客观性和准确性提出了更加严格的要求。针对当前红外成像系统最小可辨温差(minimum resolvable temperature difference, MRTD)存在的测试主观性、操作复杂性等问题, 提出基于支持向量机(support vector machine, SVM)和卷积神经网络(convolutional neural network, CNN)的两种MRTD客观测试方法。通过引入数据增强技术, 避免由训练样本少及网络层次复杂导致的过拟合。实验结果表明, 与实际人员对数据的判断相比较, MRTD测试使用SVM方法的识别准确率为94.50%, 训练时间为8.22 s; CNN方法3次训练平均准确率为99.07%, 迭代100次训练时间为487.48 s。SVM方法的实时性更好，CNN方法具有准确率高的特点，实验结果验证了这两种MRTD的客观测试方法为红外热成像系统性能指标研究提供了一种可靠的量化和评估工具。

关键词: 最小可辨温差, 机器学习, 深度学习, 支持向量机, 卷积神经网络

Abstract:

The accelerated development of infrared imaging technology has put forward more stringent requirements for the objectivity and accuracy of the testing and evaluation of infrared imaging systems. Aiming at the current problems of test subjectivity and operational complexity of the minimum resolvable temperature difference (MRTD) of infrared imaging systems, two MRTD objective test methods based on support vector machine (SVM) and convolutional neural network (CNN) are proposed. By introducing the data enhancement technique, the overfitting caused by the small training samples and the complex network hierarchy is avoided. The experimental results show that compared with the actual personnel's judgment of the data, the MRTD test using the SVM method has a recognition accuracy of 94.50% and a training time of 8.22 s, while the CNN method has an average accuracy of 99.07% in three training sessions, and a training time of 487.48 s for 100 iterations. The SVM method has better real-time performance and the CNN method is characterized by high accuracy. The experimental result verifies that these two objective test methods of MRTD provide a tool for quantification and evaluation of infrared thermal imaging system performance indicators research.

Key words: minimum resolvable temperature difference (MRTD), machine learning, deep learning, support vector machine (SVM), convolutional neural network (CNN)

中图分类号:

TP391.4

季然, 肖茂森, 李硕, 刘宇, 罗湛仪, 程嘉维. 基于机器学习的MRTD客观测试方法研究[J]. 系统工程与电子技术, 2024, 46(10): 3265-3270.

Ran JI, Maosen XIAO, Shuo LI, Yu LIU, Zhanyi LUO, Jiawei CHENG. Research on MRTD objective testing method based on machine learning[J]. Systems Engineering and Electronics, 2024, 46(10): 3265-3270.

图/表 12

表1

图1

图2

表2

图3

图4

表3

图5

图6

图7

表4

表5

参考文献 30

1	吕昕梦, 周俊. 红外成像系统MRTD测试方法研究[J]. 光电技术应用, 2021, 36 (6): 77- 80.
	LYU X M , ZHOU J . Research on MRTD testing method of infrared imaging system[J]. Electro-Optic Technology, 2021, 36 (6): 77- 80.
2	范心田, 卢振武, 孙强. 红外成像系统中MRTD测试方法的研究[J]. 红外与激光工程, 2004, 33 (5): 445- 448. doi: 10.3969/j.issn.1007-2276.2004.05.002
	FAN X T , LU Z W , SUN Q . Measurement method for MRTD of infrared imaging system[J]. Infrared and Laser Engineering, 2004, 33 (5): 445- 448. doi: 10.3969/j.issn.1007-2276.2004.05.002
3	BURROUGHS E, MOE G O, LESHER G W, et al. Automated MRTD using boundary contour system, custom feature extractors and fuzzy ARTMAP[C]//Proc. of the Infrared Imaging Systems: Design, Analysis, Modeling, and Testing, 1995, 2470: 274-287.
4	孙军月. 应用神经网络智能测量热成像系统MRTD[D]. 长春: 中国科学院长春光学精密机械与物理研究所, 2006.
	SUN J Y. Intelligent MRTD testing for thermallmaging system using ANN[D]. Changchun: Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, 2006.
5	SUN J Y, MA D. Intelligent MRTD testing for thermal imaging system using ANN[C]//Proc. of the ICO20: Remote Sensing and Infrared Devices and Systems, 2006, 6031: 260-266.
6	王连振, 吴晗平, 李旭辉, 等. 红外成像系统综合性能评价方法研究[J]. 红外技术, 2015, 37 (1): 57- 62.
	WANG L Z , WU H P , LI X H , et al. Research on the methods of IR imaging system comprehensive performance evaluation[J]. Infrared Technology, 2015, 37 (1): 57- 62.
7	LIU F Y , LIU J , WANG L B . Deep learning and infrared thermography for asphalt pavement crack severity classification[J]. Automation in Construction, 2022, 140, 104383. doi: 10.1016/j.autcon.2022.104383
8	HANG R L , LI Z , LIU Q S , et al. Hyperspectral image classification with attention-aided CNNs[J]. IEEE Trans. on Geoscience and Remote Sensing, 2021, 59 (3): 2281- 2293. doi: 10.1109/TGRS.2020.3007921
9	AHMAD M , SHABBIR S , RAZA R A , et al. Artifacts of different dimension reduction methods on hybrid CNN feature hierarchy for hyperspectral image classification[J]. Optik-International Journal for Light and Electron Optics, 2021, 246 (1): 167757.
10	ARNOUS F I , NARAYANAN R M , LI B C . Application of multidomain sensor image fusion and training data augmentation for enhanced CNN image classification[J]. Journal of Electronic Imaging, 2022, 31 (1): 013014.
11	LIU F C , XU H , QI M , et al. Depth-wise separable convolution attention module for garbage image classification[J]. Sustainability, 2022, 14 (5): 3099. doi: 10.3390/su14053099
12	MELGANI F , BRUZZONE L . Classification of hyperspectral remote sensing images with support vector machines[J]. IEEE Trans. on Geoscience and Remote Sensing, 2004, 42 (8): 1778- 1790. doi: 10.1109/TGRS.2004.831865
13	KHAIRANDISH M O , SHARMA M , JAIN V , et al. A hybrid CNN-SVM threshold segmentation approach for tumor detection and classification of MRI brain images[J]. Innovation and Research in Biomedical Engineering, 2022, 43 (4): 290- 299.
14	REDDY A S . Extensive content feature based image classification and retrieval using SVM[J]. CVR Journal of Science and Technology, 2023, 24 (1): 53- 58.
15	KEERTHANA D , VENUGOPAL V , NATH M K , et al. Hybrid convolutional neural networks with SVM classifier for classification of skin cancer[J]. Biomedical Engineering Advances, 2023, 5, 100069.
16	ARI A . Multipath feature fusion for hyperspectral image classification based on hybrid 3D/2D CNN and squeeze-excitation network[J]. Earth Science Informatics, 2023, 16 (1): 175- 191.
17	CHEN G M , CHEN Q , LONG S , et al. Quantum convolutional neural network for image classification[J]. Pattern Analysis and Applications, 2023, 26 (2): 655- 667.
18	LI B H , HOU Y T , CHE W X . Data augmentation approaches in natural language processing: a survey[J]. AI Open, 2021, 3, 71- 90.
19	POLSON N G , SCOTT S L . Data augmentation for support vector machines[J]. Bayesian Analysis, 2011, 6 (1): 1- 23.
20	田留德, 刘朝晖, 赵建科, 等. 红外热像仪MRTD测试方法研究[J]. 红外技术, 2015, 37 (5): 368- 373.
	TIAN L D , LIU C H , ZHAO J K , et al. Measurement method for MRTD of infrared imaging system[J]. Infrared Technology, 2015, 37 (5): 368- 373.
21	DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C]//Proc. of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2005: 886-893.
22	GALAR M , FERNANDEZ A , BARRENECHEA E , et al. DRCW-OVO: distance-based relative competence weighting combination for one-vs-one strategy in multi-class problems[J]. Pattern Recognition, 2015, 48 (1): 28- 42.
23	KINGMA D P , BA J . Adam: a method for stochastic optimization[J]. Computer Science, 2014, 14126980.
24	SRIVASTAVA N , HINTON G , KRIZHEVSKY A , et al. Dropout: a simple way to prevent neural networks from overfitting[J]. The Journal of Machine Learning Research, 2014, 15 (1): 1929- 1958.
25	SALEHIN I , KANG D K . A review on Dropout regularization approaches for deep neural networks within the scholarly domain[J]. Electronics, 2023, 12 (14): 3106.
26	HU J, SHEN L, SAMUEL A. Squeeze-and-excitation networks[C]//Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018: 7132-7141.
27	SARIGUL M . Performance analysis of CNN channel attention modules for image classification task[J]. Çukurova Universitesi Mühendislik Fakültesi Dergisi, 2023, 38 (1): 35- 40.
28	HSIAO T Y , CHANG Y C , CHOU H H , et al. Filter-based deep-compression with global average pooling for convolutional networks[J]. Journal of Systems Architecture, 2019, 95, 9- 18.
29	KHAN A , SOHAIL A , ZAHOORA U , et al. A survey of the recent architectures of deep convolutional neural networks[J]. Artificial Intelligence Review, 2020, 53, 5455- 5516.
30	GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2014: 580-587.

参数	取值
中波长/μm	3.7~4.8
分辨率	1 280×1 024
焦距/mm	75
F数	4.0
像素大小/μm	15×15

数据增强参数	取值
旋转范围/(°)	20
宽度偏移范围	0.1
高度偏移范围	0.1
剪切范围	0.4
变焦范围	0.5
亮度范围	[0.8, 1.2]

温差/K	模型准确率/%
温差/K	SVM	CNN
+0.20	92.10	99.98
+5.00	99.32	100.00
-0.28	88.59	99.99
-2.00	100.00	100.00

移除模块	准确率达到95%的迭代次数	测试集准确率/%
不移除任何模块	2	99.07
数据增强	5	94.00
注意力机制	4	96.00
随机失活	2	91.50

[1]	吴巍, 薛冰, 刘丹丹. 基于Tri-feature训练的目标与海杂波鉴别算法[J]. 系统工程与电子技术, 2024, 46(9): 2935-2940.
[2]	蔡伟, 王鑫, 蒋昕昊, 杨志勇, 陈栋. 基于解耦的小样本目标检测方法研究[J]. 系统工程与电子技术, 2024, 46(9): 2941-2950.
[3]	陈晓萱, 徐书文, 胡绍海, 马晓乐. 基于卷积与自注意力的红外与可见光图像融合[J]. 系统工程与电子技术, 2024, 46(8): 2641-2649.
[4]	汪强龙, 高晓光, 吴必聪, 胡子剑, 万开方. 受限玻尔兹曼机及其变体研究综述[J]. 系统工程与电子技术, 2024, 46(7): 2323-2345.
[5]	孙先涛, 江汪洋, 陈文杰, 陈伟海, 智亚丽. 基于感兴趣区域的物体抓取位姿检测[J]. 系统工程与电子技术, 2024, 46(6): 1867-1877.
[6]	王全全, 顾志豪, 吴城坤, 宛汀. 基于HHO-SVM的抗SSDF攻击协作频谱感知方法[J]. 系统工程与电子技术, 2024, 46(6): 2146-2154.
[7]	吴钇达, 王彩云, 王佳宁, 李晓飞. 基于ISVM-DS的红外多传感器融合识别方法[J]. 系统工程与电子技术, 2024, 46(5): 1555-1560.
[8]	王希, 任惠, 王威, 张嘉怡, 赵洪山. URLLC场景下信道可靠连通度预测[J]. 系统工程与电子技术, 2024, 46(5): 1810-1819.
[9]	陈雪梅, 刘志恒, 周绥平, 余航, 刘彦明. 基于HRNet的高分辨率遥感影像道路提取方法[J]. 系统工程与电子技术, 2024, 46(4): 1167-1173.
[10]	张天文, 张晓玲, 邵子康, 曾天娇. 基于掩模注意型交互的SAR舰船实例分割[J]. 系统工程与电子技术, 2024, 46(3): 831-838.
[11]	杨德贵, 许道峰. 基于时频域特征融合的IR-UWB穿墙雷达人体行为识别方法[J]. 系统工程与电子技术, 2024, 46(3): 849-858.
[12]	焦志强, 易侃, 张杰勇, 姚佩阳. 不平衡数据下基于SVM增量学习的指挥信息系统状态监控方法[J]. 系统工程与电子技术, 2024, 46(3): 992-1003.
[13]	于蕾, 邓秋月, 郑丽颖, 吴昊宇. 二阶逐层特征融合网络的图像超分辨重建[J]. 系统工程与电子技术, 2024, 46(2): 391-400.
[14]	施端阳, 林强, 胡冰, 杜小帅. 基于YOLO的航管一次雷达目标检测方法[J]. 系统工程与电子技术, 2024, 46(1): 143-151.
[15]	汪萌, 诸兵. 不确定性建模在2D和3D目标检测中的应用[J]. 系统工程与电子技术, 2023, 45(8): 2370-2376.

基于机器学习的MRTD客观测试方法研究

Research on MRTD objective testing method based on machine learning

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 30

相关文章 15

编辑推荐

Metrics

本文评价