WITHDRAWN: An accurate interpolation-based iterative frequency estimator of 2D multi- frequency sinusoids

Abstract There are many domains in which frequency estimation of two-dimensional (2D) multi-frequency sinusoids is of great concern. In this paper, based on 2D spectrum interpolation, an accurate and efficient method is proposed to estimate 2D multi-frequency in noise. The method estimates the order of the signal, adopting a coarse-fine strategy, effectively corrects frequency leakage and suppresses interference between frequency components. To obtain the coarse estimation, 2D spectrum interpolation and 2D discrete Fourier transform (2D-DFT) are adopted after the order of sinusoids has been estimated. The fine estimation, with subtraction strategies converting sinusoids into 2D single-frequency sinusoid, suppresses interference from other frequency components. Next, 2D spectrum interpolation is used to obtain an accurate frequency estimate, which accuracy is further improved through iteration. Simulation results demonstrate that the proposed method is capable of achieving the Cramer-Rao lower bound (CRLB) and outperforms other recently proposed frequency domain methods for 2D multi-frequency sinusoids.