Detecting moving targets in active sonar echograph of harbor environment using high-order time lacunarity

This paper presents a feature for detecting potential targets from high level littoral clutters in active sonar echographs. Based on lacunarity, which describes the image texture statistics, an extension to the time domain is made in order to measure the dynamic behavior of target echoes and background clutters. Moreover, as high-order moments have been shown to well characterize the non-Rayleigh tails of littoral clutter, high-order computation is incorporated in the proposed high-order time lacunarity (HOT-Lac). The potential of HOT-Lac is demonstrated using a series of active sonar echographs with diverse cooperative targets detected in real-world harbor environments in the South China Sea. Specifically, it is shown how HOT-Lac can effectively distinguish different moving small targets from high-level background clutter, and how this ability can be exploited to highlight an invasion target in harbor security.