Images from Low Frequency Active Sonar

Passive sonar systems will not be able to detect and/or classify the quiet targets of the future. Low transmit frequencies are needed to image, and hence classify, these targets at long ranges using active sonar. The low frequency sonar arrays have to be large in order to attain the angular resolution needed to image submarines at long ranges. Inhomogeneity of the propagation medium in the huge volume of water surrounding the large array creates phase errors. Additional phase errors occur because such large arrays cannot be surveyed accurately. Cost considerations dictate that the large array be thinned: this leads to high sidelobes that create image artifacts. A phenomenon called speckle causes target break-up, further reducing the likelihood of target classification. This proposed effort will develop algorithms to solve these problems associated with imaging using large and distorted low-frequency active sonar arrays. Phase errors caused by medium inhomogeneity and inaccurate element location estimates are compensated using self-cohering and self-survey algorithms. A deconvolution technique called CLEAN, initially developed in radio astronomy and later applied in coherent microwave imaging, will be used to suppress image artifacts caused by speckle and the high sidelobes of the thinned receiving array.