Underwater distorted image restoration based on Fourier wave compression sensing

Aiming at the periodic image distortion and distortion caused by the vortex formed around the bridge pier by turbulence, a compressed sensing underwater distorted image restoration algorithm based on Fourier wave was designed for AUV. Firstly, the non-sparse offset signal of pier pixels is obtained by camera sampling, and sparse transformation is carried out by using Fourier basis. Then, the sparse signal is measured by linear projection of measurement matrix and the sparse sampling network is designed to obtain high-fidelity measurement values, so as to achieve the purpose of fast dimension reduction and compression. Finally, the measured values are reconstructed and restored by the designed reconstruction network to obtain the undistorted image. Experimental results and analysis show that the proposed method can eliminate periodic image distortion and distortion, which is of great help to the automatic observation and maintenance of bridge piers.