Assessment of Collective Fish Behavior Using Imaging Sonar Under Low-Visibility Conditions

View at publisher Summarise with VIRAT
Authors: Tetsuo Imai, Ryo Harada, Kazuki Fukae, Daiki Togawa, Kenichi Arai, Hideyuki Takahashi, Hirofumi Miyajima, Mitsuru Hattori, Toru Kobayashi
Cited by: 0
References: 5

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

We evaluate imaging sonar as a practical alternative to underwater optical cameras for measuring collective fish behavior under low-visibility conditions such as nighttime and high turbidity. We reuse our prior pipeline without modification; it computes dense optical-flow features, aggregates speed and direction histograms, and classifies with a support vector machine. We apply this pipeline to imaging-sonar video recorded inside a marine fish cage. Using a day-night dataset of one-second clips, the method demonstrates robust performance under low visibility. These findings indicate that the unchanged feature design and the same SVM-based classification framework generalize across modalities from optical to acoustic video, and they support future automatic control of feeding independent of lighting and water clarity.

Fields of Study & Subjects

Water Quality Monitoring Technologies Underwater Vehicles and Communication Systems Ichthyology and Marine Biology

Publication details

Journal / source
International Conference on Consumer Electronics
Pages
1–4
Publication date
3 February 2026
Language
EN