BACKWARD CLOUD MODEL BASED FEATURE EXTRACTION OF AIRCRAFT ECHOES AND TARGET CLASSIFICATION

As a kind of complicated targets, the nonrigid vibration of aircraft, their attitude change, and the rotation of their rotating parts will induce complicated nonlinear modulation on their echoes from low-resolution radars. These kinds of modulation play an important role in target classification. However, due to the influence of clutter and noise, these kinds of modulation have the characteristics of fuzziness and randomness. As a quantitative to qualitative conversion model based on traditional probability statistics theory and fuzzy theory, backward cloud model can be used to model and analyze the modulation characteristics of the conventional low-resolution radar echoes from aircraft targets. By considering the sample values of the echo data as individual cloud droplets, the paper extracts cloud digital features such as the expectation, entropy, and hyper-entropy of each group of echo data, and investigates the application of these features in aircraft target classification based on support vector machine. The research results show that the backward cloud model can describe the aircraft echoes well, and the echo cloud digital features can be effectively used for the classification and identification of aircraft targets.