Identification Pattern Recognition Techniquefor Small-size Underwater Objects

Thepaper describes the pattern recognition technique for underwater objects based on the application of identification measurements. Identification measurements result in a number that quantitatively describes the form of any property of an informative signal. The probability of recognition and classification of underwater objects can be determined by the completeness of the pre-created database of known objects images, formed in accordance with the estimates of certified experts. With a database of three images, the probability of recognition was more than 80%. The results obtained, the high sensitivity and the absence of noticeable edge effects of the technique indicate the possibility of implementing a fundamentally new type of software for side-scan sonars, which shall improve their intellectualization in the field of recognition and classification of underwater objects.