A fast and exact seabed image reconstruction algorithm of synthetic aperture sonar

The basic principle of synthetic aperture sonar (SAS) is to record the signals received on a short physical array as it proceeds along the track, and the signals can then be combined coherently to form a long array. Conventional time delay and sum algorithm (TDS), which compensate for the echoes phase of each pixel before being summed coherently, requires too much computation to be practical in many cases. In addition, the SAS image reconstruction model is developed on the basis of uniform linear motion, and it will lead to the reconstruction image being unfocused when the motion error is more than half a quarter of the signal wavelength. However motion error leads to different phase error for different targets in the wide element beam of SAS. Compensating for phase error for each target will require considerable computations, so the exact motion compensation is hard to realize in engineering applications. A synthetic aperture signal processing method is put forward in this paper, which back propagates sonar echoes into sub-image matrixes first, and completes the synthetic aperture processing by summing of the sub-image matrixes coherently. In the method above, phase error of all targets is compensated exactly in one operation by shifting the sub-image matrixes before they are summed; and matrix summing is computed faster in modern digital signal processing servers and chips. Sea trial results show that the method can not only increase imaging speed greatly, compared with a traditional time delay and sum algorithm, but also overcomes phase error variation and precisely compensates for the motion errors. A submerged ancient town is imaged clearly by the matrix shift and sum algorithm. Both mathematical analysis and trial prove the algorithm to be very practical.