A Method for Separating Linear Scatters in Noisy Condition from High Entropy Scatters

The entropy based H-α polarimetry decomposition can be used for classifying different types of radar scatters according to their different depolarization effects: linear depolarization, random depolarization (branches, roots, random media), no depolarization (sphere, plates, and horizontal layers), and 180° depolarization (double reflections). However, the H- a feature of a linear target in the presence of strong noise with SNR < 10 dB could fall into Zone 1 (complex structures) or Zone 1 (random anisotropic scatter), and thus becoming difficult to classify. A novel approach for discriminating linear depolarization features in noisy conditions from noise-free high-entropy random polarization feature based the polarization anisotropy is proposed. The underlying theory and numerical demonstration are given and discussed.