A Novel Methodology of Estimation of Auto-Terms Domains of 2D FM Signal Representations in Noisy Environment with Application on the Wiener Space/Spatial-Frequency Filter Design

A novel methodology of estimation of the auto-terms domains of two-dimensional (2D) FM signals occupying the limited regions of frequencies in a noisy environment is considered. These domains contain full energies of the noiseless signals and hence their estimation has the key importance for the signal reconstruction. The methodology is theoretically proven and is also checked through the estimation of signal whose components occupy different ranges of frequencies and have highly nonstationary characteristics.