Related Experiment Video
Updated: Jun 8, 2025
10:56
Long-term Behavioral Tracking of Freely Swimming Weakly Electric Fish
Published on: March 6, 2014
12.5K
A comprehensive annotated image dataset for real-time fish detection in pond settings
Vijayalakshmi M1, Sasithradevi A2
1School of Electronics Engineering, Vellore Institute of Technology, Kelambakkam-vandalur road, Chennai 600127, India.
Data in Brief
|November 4, 2024
Summary
This study introduces a new fish detection dataset for aquaculture, focusing on Orange Chromide fish in Tamil Nadu ponds. The dataset aids in developing non-invasive fish monitoring systems for small-scale industries.
Area of Science:
- Aquaculture
- Computer Vision
- Marine Biology
Background:
- Fish are crucial for global food security and economies, especially in Tamil Nadu.
- Aquaculture practices require efficient, non-invasive monitoring tools for small-scale industries.
Purpose of the Study:
- To develop an efficient fish detection system for pond environments.
- To facilitate fish classification and growth monitoring in aquaculture.
- To create a valuable dataset for automated underwater image analysis.
Main Methods:
- Collected underwater video footage of Orange Chromide fish (Etroplus maculatus) in Retteri Pond, Chennai.
- Converted video to 2D images and manually annotated them using the Roboflow tool.
- Captured data under challenging conditions: occlusion, turbid water, high density, and varying light.
Main Results:
- A meticulously annotated dataset of Orange Chromide fish images.
- The dataset addresses common computer vision challenges in underwater imagery.
- Provides a resource for developing automated fish detection and monitoring tools.
Conclusions:
- The annotated dataset is a valuable resource for aquaculture, marine biology, and computer vision research.
- Enables the development of non-invasive automated systems for fish management.
- Supports advancements in small-scale aquaculture through enhanced monitoring capabilities.
