A Novel DFT-Based Algorithm for 2-D Multiple Sinusoidal Frequency Estimation

Frequency estimation of a two dimensional (2-D) multi-component sinusoidal signal in the presence of additive white Gaussian noise (AWGN) is a significant problem in various disciplines such as signal processing, radar/sonar, and wireless communications. This letter presents a novel, fast and accurate DFT-based algorithm for the frequency estimation of 2-D multi-component sinusoidal signals. We show that the proposed method attains the Cramér-Rao bound (CRB) when the parameters DFT-shift, iteration number, and minimum DFT frequency separations are chosen appropriately. Comprehensive numerical simulation results show that our algorithm almost reaches the CRB limit after a certain signal-to-noise ratio (SNR) threshold value.