Algorithm for target recognition based on fractal Brownian motion model

In the research field of target recognition, it is difficult but very important to overcome the effect which is due to the variance of the posture of the target. In other words, a model which is insensible to the posture of the target will not be found easily. In this paper, for wideband polarization radar, using the data of a plane target, the echo polarization states of the target are projected onto the polarization state plane described by the ellipticity ε and the angle τ of the polarization ellipse. We found that the statistical model of the ε and τ parameters obey the normal distribution. From this conclusion, an algorithm for target recognition based on fractal Brownian motion is proposed. From the algorithm, we can get the fractal features: fractal dimension and fractal error, and achieve 98% correct recognition rate