Centralised aquaculture farm management with AI

Intensification of aquaculture has been a driver to the rapid aquaculture growth (nearly 10 percent annually) in Asia in the past two decades, which has significantly contributed to food security and nutrition of the people and overall economy in the region. Being the most populous region of the world with heavy demands on natural resources, Asian aquaculture will face great challenges to maintain its growth, to meet the increasing demand for fish inside and outside the region over the coming decades.

Research Components

1. A knowledge driven approach to pick up the early signals of fish health conditions by combining the expert domain knowledge, physical properties and images and video processing. The scope of the prototype includes the following:

  • Domain knowledge acquisition (i.e. healthy vs unhealthy and traces of diseases)
  • Image acquisition and pre-processing
  • AI models to identify healthy and unhealthy fish
  • AI models to trace and identify disease

 

2. Develop a multispectral imaging sensor to capture a small number of spectral bands through the use of filters and illumination, based on the needs for anomaly detection. The parameters for the model to detect will be based on:

  • Visible signs on the appearance of cultured fish skin
  • Swimming behaviour as an indication to their health status
  • Monitoring the numbers and sizes of fishes in each tank

 

3. Smart Farming Tool for Data Analytics in Aquaculture

The tool collects and analyses real-time data from the farm, such as water quality, feed management, and animal health. The data is then converted into digestible insights in a mobile-based dashboard with timely alerts, which empowers farmers to follow-up with appropriate actions and solutions through machine learning.

Our partners are Amazon Web Services (AWS), Temasek Polytechnic’s Aquaculture Innovation Centre, & A*STAR.

The project in whole will take into account the common issues faced during the culture through machine learning in order to provide valuable production insights and key action items for farmers to make informed decisions for the cultured system.

SHARE THIS ARTICLE