| Popis: |
High resolution synthetic aperture sonar (SAS) imaging of small targets is easy to be confused and difficult to be identified in the background, due to the strong influence of the seabed reverberation. In this paper, an anti-submarine reverberation technology which consists of azimuth compressive sensing is used to image small targets in the seabed under the condition of sparse space. In synthetic aperture sonar, the coupling problem of two-dimensional space caused by carrier movement cannot be ignored. To overcome this problem, non-linear interpolation is adopted to achieve azimuth-range decoupling in range-Doppler domain. At the same time, the image pixels at different distances are not consistent with their azimuth, which is caused by two-dimensional decoupling. However, this paper proposes a spatial resampling technology based on compressive sensing in observation space to reconstruct the spatial spectrum of observation at different distances, so as to complete the construction of the compressive sensing observation matrix. The scattering coefficient is reconstructed by the optimist matching pursuit method to finally obtain the precise positioning imaging of SAS small target. In actual data processing, the new method in this paper can clearly imagine small target on the seabed, effectively reducing the inference of reverberation on the target imaging, and improve the recognition rate of SAS small target. |
