Using AI for Fish Biomass Estimation to Enhance Sustainable Aquaculture

Background

Sustainable Development Goal 14 (Goal 14 or SDG 14), focused on “Life Below Water,” emphasizes conserving and sustainably utilizing oceans and marine resources. To address these goals, Blue Planet Ecosystems GmbH develops automated, modular aquaculture systems to reduce reliance on fishing and traditional agriculture, alleviating stress on marine ecosystems. Accurate fish biomass estimation is crucial for monitoring animal well-being and managing sustainable aquaculture systems.

Learn more about Blue Planet:

Objective

The project aimed to create a machine learning algorithm capable of estimating fish volume and weight from video or image data. This included developing a calibration procedure for the camera-based system and exploring innovative approaches like stereovision.

Approach

Omdena collaborated with Blue Planet Ecosystems GmbH to design and implement a camera-based system for fish biomass estimation. The team utilized:

• Stereovision techniques to estimate fish volume.
• Computer vision algorithms to analyze fish dimensions and movement.
• Machine learning models to predict fish weight accurately.
• A comprehensive calibration procedure was developed to optimize system performance. Data collection included high-quality imagery and video datasets of fish in controlled environments, ensuring diverse and accurate training data.

Results and Impact

The project successfully developed an AI-based solution capable of estimating fish biomass with high precision, enabling the automation of sustainable aquaculture systems. Key outcomes included:

• Enhanced accuracy in fish weight and volume estimation.
• Reduction in manual monitoring, improving efficiency.
• Scalability for use in diverse aquaculture settings
  

This innovation supports sustainable practices, contributing to SDG 14 by minimizing overfishing and promoting eco-friendly aquaculture.

Future Implications

The findings pave the way for further advancements in automated aquaculture systems. The developed technology could influence policies promoting sustainable fisheries and aquaculture. Future research may integrate additional data sources, such as environmental parameters, to refine AI models and extend their application to other aquatic ecosystems.