The problem of inversion of convolution with the echo signal point source function is considered with the use of the regularization and maximum entropy method and further reconstruction of two-dimensional images by the method of projection in the spectral domain. The inverse convolution problem is solved for the complex-valued signal that is obtained from the real valued signal through the Hilbert transform. Numerical and experimental simulation is performed. A possibility of enhancing the echo signal along the ray’s resolution and of lowering the spectrum’s noise level with the use of complex signals (pseudo-random sequences) is demonstrated. The results are compared with those obtained using the autoregression method and the reference hologram method.