一种监督局部保持投影的飞机目标识别方法  

A Target Recognition Method for Aircraft Based on Supervised Locality Preserving Projection

作　　者：尚怡君[1] 师晓利[1] 王靳辉[2] 

机构地区：[1]郑州大学西亚斯国际学院,河南新郑451150 [2]信息工程大学,河南郑州450001

出　　处：《测绘科学技术学报》2014年第3期240-243,共4页Journal of Geomatics Science and Technology

基　　金：河南省软科学研究项目(142300410309;142400410853);郑州市科技攻关计划项目(131PPTGG426)

摘　　要：飞机目标识别方法研究在军事上具有重要意义。针对飞机目标图像数据的复杂性,提出了基于监督局部保持投影(SLPP)算法的一种目标识别方法。首先,利用PCA对飞机图像数据进行预降维,由此克服小样本问题;其次利用SLPP算法对飞机图像进行维数约简;最后,采用最近邻分类器进行飞机类型分类。通过在真实飞机图像数据库上的实验结果表明,该方法有效地提高了飞机识别的正确率。利用该方法能够有效地进行飞机目标识别。Studying the methods of the aircraft target recognition is important in military. As for the complexity of the aircraft image data, a aircraft target recognition method is presented based on the supervised locality preserving projection （ SLPP） iterative learning algorithm. Firstly, the data of the aircraft image data are reduced by principle component analysis （ PCA） in order to overcome the small-sample-size problem. Secondly, dimensionality reduc-tion is implemented by SLPP algorithm. Finally, the aircraft targets are classified by nearest neighbor classifier. The experimental results on the real aircraft image database show the effectiveness of the proposed method for air-craft target recognition. The aircraft target is effectively recognized by the proposed method.

关 键 词：飞机目标识别 维数约简 监督局部保持投影 迭代学习 最近邻分类器 

分 类 号：P237[天文地球—摄影测量与遥感] TP391[天文地球—测绘科学与技术]