An evolutionary approach for 3D superresolution imagery

The paper presents an evolutionary approach for 3D superresolution (SR) imagery combining the CLEAN method and an optimization procedure based on genetic algorithms (GA). Actually, this is an extension of some previously published research works on evolutionary programming (EP) based CLEAN method in 1D and 2D cases. Measured data results obtained using GA-based CLEAN-1D and CLEAN-2D are first provided in the paper. A gap is thus filled since previous works use only synthetic generated ISAR data. For the 3D case, the main idea is still to consider the reconstruction process as an optimization problem related to the residual energy of the acquired data after each scattering center (SC) extraction and cancellation. However, a new spatial dimension is added. The proposed solution takes advantage of some powerful convergence properties of GA and provides good performance in terms of both accuracy and robustness.