摘要
对目标识别中的特征提取技术进行了研究,介绍了两种用于被动声信号特征提取的方法,分别将谐波集(HS)频率和不同尺度小波子空间能量作为特征矢量,给出相应的算法;并利用实测信号将这些技术运用到直升机目标的识别问题中,利用 kNN 分类器对直升机目标和非直升机目标进行分类。结果表明这两种方法都能达到较高的正确识别率。
The paper discusses the feature extraction technologies in the target classification, and introduces two ways to extract features from the passive acoustic signals. These ways employ, respectively, the harmonic sets frequencies and the energies in different scales after the wavelet decomposition as the feature vectors. Also the simulation is presented using the real data and the kNN classifier to identify the helicopter. The result shows that these two technologies can achieve high accuracy in the helicopter classification problem.
出处
《探测与控制学报》
CSCD
北大核心
2004年第4期1-4,共4页
Journal of Detection & Control
基金
国防重点实验室基金资助项目(51454020101HK0307)
西北工业大学创新基金资助(2003CR080001)
