Pre-processing Algorithm for Target Classification Based on Energy Distribution Characteristics in Vertical Multi-Beam Forward-Looking SONAR and Its Sea Trial Validation

This paper proposes pre-processing algorithm that can effectively distinguish between extended and small targets based on the target energy distribution characteristics of 3D SONAR(SOund Navigation And Ranging) data. 3D SONAR data is obtained by applying signal processing techniques to multi-channel data acquired through vertical multi-beam FLS(Forward-Looking SONAR). To improve the classification performance of the energy distribution data, image pre-processing filters for extended and small targets are designed respectively. Each filter is applied to the same energy distribution data to verify that the resulting image achieves distinct classification based on target size. Simulation results validate the proposed method, and actual sea trials confirm its effectiveness.