Aerial target classification based on feature analysis

Abstract

Aircrafts are widespread used in the modern warfare. A radar therefore should have the capability to classify and identify the aerial targets. In this study, a method based on feature analysis is proposed to classify and identify the aircraft targets. First, multiple features are extracted and analysed for further utilisation, including Doppler shift, relative amplitude, waveform entropy of time and frequency domain, and time–frequency characteristics. Second, the features are employed for classification purpose by means of the support vector machine. Moreover, a method is presented to identify the helicopter with odd or even number of blades, owing to the width of time–frequency distributions of the returned signals. The performance of the discussed method is verified with the experimental results of measured data.

Keywords

• image classification
• target tracking
• support vector machines
• signal classification
• pattern classification
• helicopters
• object detection
• time-frequency analysis
• entropy
• feature extraction
• Doppler shift
• aerial target classification
• feature analysis
• aircrafts
• modern warfare
• aerial targets
• aircraft targets
• multiple features
• Doppler shift
• relative amplitude
• waveform entropy
• frequency domain
• time–frequency characteristics
• classification purpose
• support vector machine
• time–frequency distributions
• discussed method