Motion Acoustic Flow Field: Motion Estimation for Blob Targets in Active Sonar Echograph of Harbor Environments

TL;DR

This paper introduces a novel motion acoustic flow field (MAFF) method to accurately estimate the motion of blob targets in active sonar echographs within harbor environments, overcoming noise and interference challenges.

Contribution

It proposes a new MAFF function and solving strategy tailored for sonar echographs, addressing limitations of traditional optical flow methods in noisy harbor conditions.

Findings

MAFF effectively captures subtle target motions in real-world harbor sonar data.

The method outperforms traditional optical flow approaches under noisy conditions.

Experimental results validate the robustness and accuracy of MAFF.

Abstract

Motion feature is of great significance for blob targets recognition, behavior analysis and threat estimation in active sonar echographs. Hence, it is desirable to access the space-time variation of echo intensity on each spatial-temporal resolution cell from the sonar echographs sequence. Then the subtle motion information of the potential blob targets can be accurately characterized. This idea has been conduced in optical image sequences by solving an motion optical flow field (MOFF) function. Nonetheless, due to the sparkle of the sonar echograph sequences, and strong interferences caused by wake and cavitation noise of fast-moving ship in harbor environments, the constraints underlying the traditional motion optical flow function that is couples the brightness constancy constant along time dimension of each echo intensity points and the motion field spatial smoothness does not hold in our case. Hence, this paper presents a new motion acoustic flow field (MAFF) function and its solving strategy to accurately characterize the subtle motion information of blob targets in active sonar echographs of harbor environments. Experiments on a series of cooperative targets in real-world harbor environments demonstrate the efficacy of our proposed MAFF.