Modified CLEAN Algorithm Based on Two-Dimensional Matched Filter Outputs

The CLEAN algorithm is usually used to reconstruct images due to its low computational complexity. However, existing CLEAN algorithms cannot completely remove signal components of strong targets due to the low accuracy in estimating the target amplitudes, which results in weak targets being masked by the residual range-Doppler sidelobes generated by strong targets. In this paper, a modified CLEAN algorithm based on two-dimensional matched filter outputs (Modified MF-CLEAN) is proposed. Modified MF-CLEAN can improve the target amplitude estimation accuracy by performing multiple corrections to the amplitude estimation during the iterations. Simulation results show that Modified MF-CLEAN significantly outperforms the existing CLEAN algorithms in terms of the target amplitude estimation performance and success rate.