A fractal approach to radar target recognition

A method of radar target classification and recognition through extracting their fractal features directly from the radar echoes is presented. Firstly, two dimensional texture images with distances and directions as the coordinate axes are formed, then, the FFT-Mellin transformation is applied to eliminate the affect caused by flights' motion, posture and direction. Finally, a fractal texture analysis method is used to extract the features of the planes, which is improved in several points: to estimate the fractal dimensions in different directions, to calculate the fractal dimensions by using divisional fitting methods and to use the fractal signatures as the feature vectors. The results show that our method performs well in radar target recognition.