Mississippi State University Harnessing Machine Learning to Estimate Aquaculture Production and Value Chain Performance in Bangladesh Annual Report April to September 2020
Views
0% 0
Downloads
0 0%
Timeless limited access
The Harnessing Machine Learning for Better Aquaculture project began operations on April 1, 2020.
The project is implemented by Michigan State University (MSU) in partnership with Bangladesh
Agricultural University (BAU), WorldFish (Bangladesh office) and the International Center for
Tropical Agriculture (CIAT). The project has three goals
First, identify emerging technologies and innovative practices in aquaculture value chains and pilot
digital extension approaches that accelerate their adoption. Second, use machine learning to
automate extraction of data on ponds from satellite images and integrate with georeferenced survey
data to accurately estimate fish production, economic value, and employment. Third, build
organizational and individual capacity in Bangladesh for conducting rigorous research on socioeconomic
and spatial dimensions of aquaculture.
The project is comprised of three components that feed into these two goals: (1) Surveys; (2)
Remote sensing; (3) Capacity building. Component 1 will survey a sample of 1100 hatcheries, feed
suppliers, farmers, and fish traders. Component 2 will utilize machine learning to extract and
analyze data on fishponds from satellite images to facilitate development of an interactive online
data visualization tool utilizing data from component 1. Component 3 is dedicated to formal training
and outreach that builds individual, organizational and societal capacity.
Good progress was made during the first two quarters of the project (1 April 2020 - 30 September 2020) towards activities under
components 1 and 2. Activities oriented to component 3 will be initiated later on in the project cycle.
Citation
Ben Belton. (30/9/2020). Mississippi State University Harnessing Machine Learning to Estimate Aquaculture Production and Value Chain Performance in Bangladesh Annual Report April to September 2020.