Overview of underwater target feature extraction methods

Abstract

Target of underwater environments classification and distinguish using active sonar is very complex signal processing work, therefore, many research institutions in the world study the classification and recognition methods of underwater small targets using active sonar.The recognition methods of static underwater small targets can be roughly divided into imaging method and echo method. The imaging method is based on high-frequency narrow beam to obtain the acoustic image of the target and its shadow while suppressing the reverberation interference, so as to realize the classification and recognition of the target. Imaging sonar has high working frequency, limited sonar range, and can not effectively identify false targets with similar geometry. Bright spot model and resonance scattering theory are important theoretical basis of target echo method. The two theories agree that the echo is the result of the interaction between the incident signal and the underwater target, which must contain the characteristics of the target, so it is feasible to detect the target by analyzing the characteristics of the echo signal. However, in the actual underwater target detection process, due to the weak target echo signal, complex seabed environment and target geometry, uncertain sound propagation conditions and so on, the target recognition based on echo method still has great challenges.