Low Complexity Orthogonal Matching Pursuit Algorithm for High-Resolution Imaging Sonar

We propose a segment focusing compressed sensing algorithm which is high-resolution, low complexity and uses inherent sparsity in the scene of sonar imaging. We demonstrate on simulation and through poor experimental data that Fast Fourier Transform (FFT) beam forming and the proposed algorithm can distinguish the targets generated by the direct path target scattering echo signal. But FFT beam forming has a wider main lobe and high side lobe level. When the main lobe amplitude is 0dB, the maximum side lobe amplitude is -3.3dB. Minimum Variance Distortionless Response (MVDR) is unable to distinguish the two targets and has wider main lobe and high side lobe level. When the main lobe amplitude is 0dB, the maximum side lobe amplitude is -4.2dB. The proposed algorithm can get a sparse solution. FFT beam forming and MVDR are unable to distinguish the two targets generated by a multipath scattering echo signal. In contrast, the proposed algorithm could effectively distinguish the two targets generated by multipath scattering echo signal.