Aircraft classification method based on the kurtosis–skewness feature and wavelet decomposition and linear discriminant analysis

Abstract

At present, most of the jet engine modulation feature extraction methods are based on the modulation wave period or the inter-spectral interval of the modulation line spectrum. However, such spectral estimation methods are often difficult to obtain good classification performance due to the signal-to-noise ratio, pulse repetition frequency (PRF) and observation time. The statistical analysis of the three types of aircraft target echoes shows that there is a significant difference in the normalised amplitude distribution, and based on this, the kurtosis–skewness feature is extracted to classify the targets. This feature has a strong anti-noise capability, the requirement for PRF and observation time is not high, and one of the parameters can be used to make up for another parameter, so we can make a balance between PRF and time if needed. The simulation test proves that the proposed method has good classification performance.

Keywords

• statistical analysis
• spectral analysis
• noise
• feature extraction
• jet engines
• wavelet transforms
• modulation
• image classification
• pattern classification
• linear discriminant analysis
• jet engine modulation feature
• extraction methods
• modulation wave period
• inter-spectral interval
• modulation line spectrum
• spectral estimation methods
• signal-to-noise ratio
• prf
• observation time
• statistical analysis
• aircraft target echoes
• normalised amplitude distribution
• kurtosis
• skewness feature
• strong anti-noise capability
• aircraft classification method
• wavelet decomposition