针对不变矩对仿射形变目标描述的不足，为提高舰船型号的识别精度，提出一种基于小波和仿射不变矩特征融合的舰船型号识别方法。首先对二值舰船图像进行归一化处理，并分别提取归一化舰船图像的小波矩特征值和仿射不变矩特征值；然后通过计算样本特征均值与标准差的比值，选择出鲁棒性好、稳定性高的特征，通过归一化方法进行融合；最后构造五类舰船的样本集，采用支持向量机（SVM）作为分类器识别测试样本的型号，分析不同矩特征、样本集大小、SVM参数、本文方法对识别精度、稳定性的影响。实验结果表明，文中给出的算法提高了识别精度，并且在训练样本集较小时仍能获得88%以上的识别率。
According to the shortage of invariant moment in deformation object description, in order to improve the recognition accuracy of warship type, this paper proposes a new warship type identification method. First of all, according to the normalized binary images of warship, the features of wavelet moment and affine invariant moment are extracted respectively; Then, the features with good robustness and high stability are selected by calculating the ratio of mean and standard deviation of the sample features, and two different features are fused by normalization method in order to eliminate the differences between two features; Finally, five types of sample set of the warships are constructed through Matlab program, the support vector machine (SVM) is used as classifier to identify the warship type of test sample set which consists of the whole sample set except the training set, and the diffenrences among wavelet moment, affine invariant moment and the proposed method are compared in recognition accuracy and the effect of the training sample set size and the parameters of SVM on the identification accuracy. Experimental results show that the proposed algorithm improves the recognition accuracy, and the recognition accuracy is still greater than 88% when the training sample set is small.
2017,39(8): 170-175 收稿日期：2016-07-22
DOI：10.3404/j.issn.1672-7649.2017.08.036
分类号：TP319.4
基金项目：江苏省高校自然科学研究项目（15KJB510008）；毫米波国家重点实验室开放课题（K201714）
作者简介：陈慧珺(1991-)，女，硕士研究生，研究方向为图像处理
参考文献：
[1] 李平, 魏仲慧, 何昕, 等.采用径向Tchebichef矩不变量的飞机型号识别[J].光电子·激光, 2014, 25(2): 364-371.LI Ping, WEI Zhong-hui, HE Xin, et al.Aircraft type recognition using radial-tchebichef moment invariant[J].Journal of Optoelectronics·Laser, 2014, 25(2): 364-371.
[2] 王占领, 张登福, 王世强.雷达辐射源信号的伪Zernike矩双谱二次特征提取[J].现代雷达, 2013, 35(11): 39-41.WANG Zhan-ling, ZHANG Deng-fu, WANG Shi-qiang.Rader emitter signal bispectrum cascade feature extraction based on pseudo-Zernike moment[J].Modern Rader, 2013, 35(11): 39-41.
[3] 孙贝.基于矩特征提取的图像识别算法研究[D].太原: 太原理工大学, 2010.
[4] SHEN D, HORACE H S Ip.Discriminative wavelet shape descriptors forrecognition of 2-D patterns[J].Pattern Recognition, 1999, 32(2): 151-165.
[5] 薛鹏飞, 胡玉兰.小波和仿射不变矩的目标特征提取方法研究[J].沈阳理工大学学报, 2012, 31(3): 11-14.XUE Peng-fei, HU Yu-lan.Method of feature extraction of the target based on wavelet moment and affine invariant moment[J].Journal of Shenyang Ligong University, 2012, 31(3): 11-14.
[6] 于吉红, 吕俊伟, 白晓明.一种基于组合不变矩的新的舰船图像目标识别方法[J].红外, 2011, 32(9): 23-28.YU Ji-hong, LV Jun-wei, BAI Xiao-ming.A new method for ship image target recognition based on combined invariant moments[J].Infraded, 2011, 32(9): 23-28.
[7] 付俊.基于特征融合的遥感图像飞机目标识别研究[D].南昌, 南昌航空大学, 2014: 1-2.
[8] 刘正君, 李琦, 王骐.基于组合矩的激光雷达距离像目标识别[J].中国激光, 2012, 39(6): 1-7.LIU Zhen-jun, LI Qi, WANG Qi.Object recognition of ladar range image using combined moment invariants[J].Chinese Journal of Lasers, 2012, 39(6): 1-7.
[9] 戴娟, 顾斌杰, 潘丰.基于引力搜索算法的SVM参数优化及应用[J].江南大学学报(自然科学版), 2013, 12(2): 127-131.DAN Juan, GU Bin-jie, PAN Feng.Parameter optimization and application of SVM based on gravitational search algorithm[J].Journal of Jiangnan University(Natural Science Edition), 2013, 12(2): 127-131.
[10] 叶福玲.基于小波矩和叶形特征的叶片识别[J].福建工程学院学报, 2014, 12(1): 79-82.YE Fu-ling.Leaf recognition based on wavelet moment and the features of leaf shape[J].Journal of Fujian University of Technology, 2014, 12(1): 79-82.
[11] JAN F, TOMAS S.A moment-based approach to registration of images with affine geometric distrotion[J].IEEE Transactions on Geoscience and Remote Sensing, 1994, 32(2): 382-387.
[12] 王小川, 史峰, 郁磊, 等.Matlab神经网络43个案例分析[M].北京: 北京航空航天大学出版社, 2013: 133-134.WANG Xiao-chuan, SHI Feng, YU Lei, et al.Matlab neural network with 43 cases analysis[M].Beijing: Beihang University Press, 2013: 133-134.
