Optimization of technical characteristics of radio vision systems is considered in the radars with ultra-wideband sounding signals. Highly noisy conditions, in which such systems operate, determine the requirements that should be met by the signals being studied. The presence of the multiplicative noise makes it difficult to design optimal algorithms of echo-signal processing. Consideration is being given to the problem of discriminating objects hidden under upper layers of the ground at depths comparable to the probing pulse duration. Based upon the cepstrum and textural analysis, a subsurface radar signal processing technique has been suggested. It is shown that, however the shape of the probing signal spectrum might be, the responses from point targets in the cepstrum images of subsurface ground layers make up the texture whose distinctive features enable objects to be detected and identified.