Classification and Recognition Algorithm for Long-wave Infrared Targets Based on Support Vector Machine
-
摘要: 红外图像的分辨率低和色彩单一,但由于红外设备的全天候工作特点,因而在某些场景具有重要作用。本文采用一种基于支持向量机(support vector machine, SVM)的长波红外目标图像分类识别的算法,在一幅图像中,将算法提取的边缘特征和纹理特征作为目标的识别特征,输入到支持向量机,最后输出目标的类别。在实验中,设计方向梯度直方图+灰度共生矩阵+支持向量机的组合算法模型,采集8种人物目标场景图像进行训练和测试,实验结果显示:相同或者不相同人物目标,穿着不同服饰,算法模型的分类识别正确率较高。因此,在安防监控、工业检测、军事目标识别等运用领域,此组合算法模型可以满足需要,在红外目标识别领域具有一定的优越性。Abstract: Infrared images have a low resolution and a single color, but they play an important role in some scenes because they can be used under all weather conditions. This study adopts a support vector machine algorithm for long-wave infrared target image classification and recognition. The algorithm extracts edge and texture features, which are used as the recognition features of the target, and forwards them to a support vector machine. Then, the target category is output for infrared target recognition. Several models, such as the histogram of oriented gradient, gray level co-occurrence matrix, and support vector machine, are combined to collect images of eight types of target scenes for training and testing. The experimental results show that the algorithm can classify the same target person wearing different clothes with high accuracy and that it has a good classification effect on different target characters. Therefore, under certain scene conditions, this combined algorithm model can meet the needs and has certain advantages in the field of target recognition.
-
表 1 混淆矩阵
Table 1. Confusion matrix
Confusion matrix Predictivevalue Class 1 Class 2 Class 3 …. Class 8 Real value Class 1 T11 F12 F13 … F18 Class 2 F21 T22 F23 … F28 Class 3 F31 F32 T33 … F38 …. … … … … .. Class 8 F81 F82 F83 … T88 表 2 图像场景代表意义
Table 2. Image scene representative meaning
Class Representative meaning Class A Target U wears camouflage clothes indoors at night Class B Target U wears ordinary clothes indoors at night Class C Target U wears camouflage clothes outdoors during the day Class D Target U wears camouflage clothes outdoors at night Class E Target V wears ordinary clothes outdoors at night Class F Target W wears ordinary clothes indoors at night Class G Target X wears ordinary clothes outdoors during the day Class H Target Y wears ordinary clothes outdoors during the day 表 3 模型分类结果
Table 3. Model classification results
Confusion matrix Predictive value Class A Class B Class C Class D Class E Class F Class G Class G Real value Class A 111 0 0 0 0 0 0 0 Class B 1 110 0 0 0 0 0 0 Class C 0 0 43 1 2 0 1 0 Class D 0 0 3 44 0 0 0 0 Class E 0 0 2 2 31 0 0 0 Class F 0 0 1 2 2 16 0 0 Class G 0 0 7 3 6 0 4 0 Class H 0 0 2 2 1 1 0 12 -
[1] 曹凤杰. 红外图像人脸识别方法研究[D]. 西安: 西安电子科技大学, 2010.CAO Fengjie. Research on Infrared Image Face Recognition Method[D]. Xi'an: Xidian University, 2010. [2] Der S Z, Chellappa R. Probe-based automatic target recognition in infrared imagery[J]. IEEE Transactions on Image Processing, 1997, 6(1): 92-102. doi: 10.1109/83.552099 [3] 姜锦锋. 红外图像的目标检测、识别与跟踪技术研究[D]. 西安: 西北工业大学, 2004.JIANG Jinfeng. Research on Target Detection, Recognition and Tracking Technology of Infrared Image[D]. Xi'an: Northwestern Polytechnical University, 2004. [4] 郭济民. 基于深度神经网络的物体识别方法研究及实现[D]. 成都: 电子科技大学, 2018.GUO Jimin. Research and Implementation of Object Recognition Method Based on Deep Neural Network[D]. Chengdu: University of Electronic Science and Technology, 2018. [5] Abdulkadir Eryildirim, Ibrahim Onaran. Pulse Doppler radar target recognition using a two-stage SVM procedure[J]. IEEE Transactions on Aerospace and Electronic Systems, 2011, 47(2): 1450-1457. doi: 10.1109/TAES.2011.5751269 [6] 李小迷. 葡萄糖药液中异物目标视觉检测与识别方法研究[D]. 长沙: 湖南大学, 2010.LI Xiaomi. Research on Visual Inspection and Recognition Method of Foreign Objects in Glucose Liquid[D]. Changsha: Hunan University, 2010. [7] 王朔琛, 汪西莉, 马君亮. 基于均值漂移的半监督支持向量机图像分类[J]. 计算机应用, 2014, 34(8): 2399-2403. doi: 10.3969/j.issn.1001-3695.2014.08.038WANG Shuochen, WANG Xili, MA Junliang. Semi-supervised support vector machine image classification based on mean shift[J]. Journal of Computer Applications, 2014, 34(8): 2399-2403. doi: 10.3969/j.issn.1001-3695.2014.08.038 [8] 丁方静. 室内监控中移动检测与跟踪算法的改进与实现[D]. 南京: 东南大学, 2017.DING Fangjing. Improvement and Implementation of Moving Detection and Tracking Algorithm in Indoor Monitoring[D]. Nanjing: Southeast University, 2017 [9] 卞海曼. 基于卷积神经网络的行人检测[D]. 合肥: 合肥工业大学, 2017.BIAN Haiman. Pedestrian Detection Based on Convolutional Neural Network[D]. Hefei: Hefei University of Technology, 2017. [10] Navneet Dalal, Bill Triggs. Histograms of oriented gradients for human detection[C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), 2005. [11] Minho J, Y Hee Yong, C Hsiao-Hwa. Intelligent RFID tag detection using support vector machine[J]. IEEE Transactions on Wireless Communications, 2009, 8(10): 5050-5059. doi: 10.1109/TWC.2009.071198 [12] WANG R P, CHEN J, SHAN S G, et al. Enhancing training set for face detection based on SVM[J]. Journal of Software, 2009, 19(11): 2921-2931. doi: 10.3724/SP.J.1001.2008.02921 [13] 尤倩. 基于SVM的脱机手写体数字识别的研究与应用[D]. 济南: 山东师范大学, 2014.YOU Qian. Research and Application of Offline Handwritten Digit Recognition Based on SVM[D]. Jinan: Shandong Normal University, 2014. [14] 张小琴, 赵池航, 沙月进. 基于HOG特征及支持向量机的车辆品牌识别方法[J]. 东南大学学报: 自然科学版, 2013, 43(2): 410-413. https://www.cnki.com.cn/Article/CJFDTOTAL-DNDX2013S2041.htmZHANG Xiaoqin, ZHAO Chihang, SHA Yuejin. Vehicle brand recognition method based on HOG features and support vector machine[J]. Journal of Southeast University: Natural Science Edition, 2013, 43(2): 410-413. https://www.cnki.com.cn/Article/CJFDTOTAL-DNDX2013S2041.htm [15] LI Weixing, SU Haijun, PAN Feng, et al. A fast pedestrian detection via modified HOG feature[C]//Proceedings of the 34th Chinese Control Conference of IEEE, 2015: 3870-3873. [16] 曾雪. 基于旋转不变梯度方向直方图的航拍图像目标检测[D]. 南京: 东南大学, 2017.ZENG Xue. Object Detection Based on Rotation Invariant Histogram of Oriented Gradient in Aerial Image[D]. Nanjing: Southeast University, 2017. [17] Alex Omid-Zohoor, Christopher Young, David Ta, et al. Toward always-on mobile object detection: energy versus performance tradeoffs for embedded HOG feature extraction[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2018, 28(5): 1102-1115. doi: 10.1109/TCSVT.2017.2653187 [18] 段嘉欣. 基于梯度下降的时变PID算法[J]. 中国新通信, 2019, 21(14): 223-226. https://www.cnki.com.cn/Article/CJFDTOTAL-TXWL201914177.htmDUAN Jiaxin. Time-varying PID algorithm based on gradient descent[J]. China New Telecommunications, 2019, 21(14): 223-226. https://www.cnki.com.cn/Article/CJFDTOTAL-TXWL201914177.htm [19] CHEN Pei-Yin, HUANG Chien-Chuan, Lien Chih-Yuan, et al. An efficient hardware implementation of HOG feature extraction for human detection[J]. IEEE Transactions on Intelligent Transportation Systems, 2014, 15(2): 656-662. doi: 10.1109/TITS.2013.2284666 [20] HE Jiayuan, ZHU Xiangyang. Combining improved gray-level co-occurrence matrix with high density grid for myoelectric control robustness to electrode shift[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2017, 25(9): 1539-1548. doi: 10.1109/TNSRE.2016.2644264 [21] 叶鹏, 王永芳, 夏雨蒙, 等. 一种融合深度基于灰度共生矩阵的感知模型[J]. 计算机科学, 2019, 46(3): 92-96. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJA201903012.htmYE Peng, WANG Yongfang, XIA Yumeng, et al. Perceptual model based on GLCM combined with depth[J]. Computer Science, 2019, 46(3): 92-96. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJA201903012.htm [22] 王红, 武继刚, 张铮. 基于二维MB_LBP特征的人脸识别[J]. 计算机工程与应用, 2015, 51(10): 191-194. doi: 10.3778/j.issn.1002-8331.1305-0396WANG Hong, WU Jigang, ZHANG Zheng. Face recognition based on 2-dimensional MB-LBP characteristics[J]. Computer Engineering and Applications, 2015, 51(10): 191-194. doi: 10.3778/j.issn.1002-8331.1305-0396 [23] Marceau D J, Howarth P J, Dubois J M, et al. Evaluation of the grey -level co-occurrence matrix method for land-cover classification using SPOT imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 1990, 28(4): 513-519. doi: 10.1109/TGRS.1990.572937 [24] Simon D, Simon D L. Analytic confusion matrix bounds for fault detection and isolation using a sum-of-squared-residuals approach[J]. IEEE Transactions on Reliability, 2010, 59(2): 287-296. doi: 10.1109/TR.2010.2046772