摘要
研究自适应联合时频处理的抗箔条干扰技术,利用时域、频域和联合时频域几何矩进行模式的特征提取,以充分节省计算资源,并尽可能多地保留用于目标识别的信息;其次,应用主成分分析进一步降低特征空间的维数,将得到的特征向量用多层感知神经网络进行识别.为了证明所提出模式的性能,对3种目标(含箔条云目标)进行了识别.结果表明,所提出的技术在目标识别抗箔条云干扰方面具有显著的潜在应用价值.
This paper presented a new anti-cloud jamming scheme, which uses adaptive joint time -frequency processing techniques. The exact and dosed form expressions of geometrical moments of the adaptive spectrogram(ADS) in the time, frequency, and joint time-frequeney domains were derived and features obtained by this method co provide substantial savings of computational resources and preserve as much essential information. Next, a principal eomponent analysis(PCA) was used to further reduce the dimension of feature space, and the resuhing feature vectors were passed to the classifier stage based on the multilayer perceptron neural network. To demonstrate the performance of the proposed scheme, three targets(ineluding chaff cloud) were identified. The results show that the proposed technique has a significant potential in the identification of anti-cloud jamming by classifying targets.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2008年第7期1211-1216,共6页
Journal of Shanghai Jiaotong University
基金
国防预研资助项目(51473030101JW0301)
关键词
主分量分析
自适应谱
时频处理
自适应高斯表示
principal component analysis(PCA)
adaptive spectrogram(ADS)
time-frequency processing
adaptive Gaussian representation(AGR)