TARGET CLASSIFICATION WITH LOW-RESOLUTION RADARS BASED ON MULTIFRACTAL FEATURES IN FRACTIONAL FOURIER DOMAIN

Due to the limitations of low-resolution radar system and background clutter, the task of target classification with conventional low-resolution radars is relatively difficult. This paper introduces fractional Fourier transform (FrFT) to process aircraft echoes in order to find the optimal fractional Fourier domain, in which signal to noise ratio can reach the maximum, and then applies multifractal theory to the feature extraction of radar targets. Based on the above, we use SVM to do target classification. Experiments show that the multifractal characteristics of aircraft echoes can be enhanced by FrFT, and the features extracted from the optimal fractional Fourier domain can be used effectively to classify different types of aircraft even in the case of low SNR.