Landscape Recognition Algorithm for Polarized Detector based on Optimal Projection

In this paper, a novel landscape recognition algorithm based on the optimal projection is proposed to meet the landscape measurement requirements of a detector. The proposed algorithm is applied to the airborne radar platform, proving a necessary recognition ability. The characteristic values of the target scattering matrix eigenvalues are defined, which has no direct relation to the power and angle of a detection signal and is only related to the topographic features of a landscape in the detection range. Besides, a series of standard scatter rotation invariant matrices with uncertain parameters are constructed based on the Krogager's idea of decomposing the target scattering matrices. The geomorphological polarization scattering matrix of a detection range is projected onto the standard scatter projects, and relevant weight values are calculated to judge the landscape attributes so as to achieve the goal of landscape recognition. Effective landscape recognition can further enhance the autonomous ability of ammunition and overcome the effects of a new diversified operational environment.