Discrimination and suppression of reverberation by leveraging the property disparity between target echoes and rough seabed reverberation

In shallow water environments, reverberation, which is generated from the rough seabed, remains one of the major sources of background interference in active sonar systems. This paper proposes a reverberation suppression method that leverages multidimensional property differences, designed for a rough seabed. The approach establishes a link between variations in seabed reflection and scattering coefficients and the spectral property of reverberation signals, providing a physical basis for effective suppression. Using the Wigner–Ville distribution, reverberation suppression is formulated as the identification and utilization of time–frequency feature disparities between reverberation and target echoes. The method consists of two stages. First, seed region growing is enhanced using a multi-feature framework and Rényi entropy to suppress reverberation in the time–frequency plane. The multi-feature framework evaluates the stability and similarity of the instantaneous frequency sequence to screen the initial target echo position, whereas the Rényi entropy of the time–frequency domain adaptively adjusts the growing thresholding. In the second stage, the extracted time–frequency results are combined with a de-chirp procedure to reconstruct the time-domain echoes. Simulation and experimental results under various shallow water conditions demonstrate that the proposed method effectively suppresses reverberation while accurately preserving the target echo.