Announcing Our Inaugural Round of Funding for Agricultural Datasets for AI

We are thrilled to share Lacuna Fund’s first cohort of supported projects in the agricultural AI for social good domain. With over 100 applications from, or in partnership with, organizations across Africa, we were deeply encouraged by the depth and breadth of the proposals.

The recipients of this first round of funding are unlocking the power of machine learning to alleviate food security challenges, spur economic opportunities, and give researchers, farmers, communities, and policymakers access to superior agricultural datasets. We are proud to support their work.

Decision Support Tool for Community-led Land Use Plans

Livestock play a crucial role in poverty alleviation, particularly for the rural poor in sub-Saharan Africa who depend on animals for survival. In this part of Africa, the majority of livestock are raised in traditional systems that rely on the sharing of communal resource areas, such as pastures and water.

This project, which is led by The Nelson-Mandela African Institution of Science and Technology, University of Glasgow, University of Hohenheim, and NOTTECH Company Limited, will apply participatory methods and telemetry data to generate labeled datasets of landscape features. These labeled datasets will then be used to develop a community-led land use plan in Tanzania to support the management of livestock resources and reduce conflicts with crop cultivation.

Land use management and resource sharing is very vital in land based economies. This becomes even more apparent where livestock are a major part of the crop-livestock mixed systems. We are very excited and optimistic to contribute to the land use planning process through a data-driven and evidence-based tool.”

The Nelson-Mandela African Institution of Science and Technology

Eyes on the Ground: Providing Quality Model Training Data through Smartphones

In the Eyes on the Ground project, the team from ACRE Africa and the International Food Policy Research Institute (IFPRI) will use smartphones to create a unique dataset of georeferenced crop images along with labels on input use, crop management, phenology, crop damage, and yields. Images will be collected in 11 counties in Kenya. Funding will be used to support data collection, standardization of methods for submitting images, and curation of datasets so they meet best practices for ground reference data collection and cataloguing, while safeguarding ethics considerations.

This is a novel concept that endeavors to provide smallholder farmers with risk mitigation and adaptation strategies through satellites and smart phones to ensure that they invest in high productivity agriculture investments. The ground pictures not only provide ACRE Africa with Eyes on the Ground to fine-tune insurance products/models and minimize basis risk, but also enable us to observe management practices that promote the adoption of productivity-enhancing yet resilient technologies through bundling with stress-tolerant seeds and remote advisories. To bridge the gap between insurance products, resilient technologies and smallholder farmers we leverage on the Village Extension Service Providers (VESPs) model, a high touch and robust route to market model to help create awareness, capacitate and distribute our products. This model ultimately intends to create entrepreneurial opportunities.”

Lilly Waithaka
ACRE

Helmets Labeling Crops

This project will create unprecedented ML-ready labeled datasets for crop type, crop pest and disease, and market prices in the main food production regions in five African countries. This project is a collaborative initiative between the University of Maryland/NASA Harvest Africa Program, The Regional Universities Forum for Capacity Building in Agriculture (RUFORUM), Center for Earth Observation and Citizen Science at the International Institute for Applied Systems Analysis, The Regional Center for Mapping of Resource for Development (RCMRD), The Eastern Africa Grain Council (EAGC), Lutheran World Relief Mali, International Maize and Wheat Improvement Center (CIMMYT), and Radiant Earth Foundation.

Collectively, the team will use novel and innovative approaches that include rapid point data collection with cameras mounted on the hoods of vehicles—“helmets”—combined with crowdsourcing to create point and polygon labels. By partnering with local universities, this project will create opportunities for training future African researchers to use remote sensing and machine learning.

My favorite part of my work is to spend time in the field. There’s so much to learn from the ground about the people, about the crops, and understanding why it is all-important. “Helmets Labeling Crops” is one of a kind opportunity to address a critical gap, to learn and develop a cost-effective scalable approach to labeled data needed to improve the basis of agricultural monitoring. We’ll also get a chance to train and work with students across Africa which is an exciting opportunity to encourage students into remote sensing, data science, and machine learning.”

Catherine Lilian Nakalembe
University of Maryland

IoT Water Quality Monitoring System for Freshwater Aquaponics Fish Ponds

Many in Sub-Saharan Africa live below the poverty line, leading to low protein intake. Aquaculture, the science of breeding fish for local protein consumption and commercial ventures, presents one solution. The University of Nigeria Nsukka will build a remotely monitored and controlled Internet of Things (IoT) water quality management system for conventional ponds and the aquaponics fish pond systems to generate labeled datasets, and partner with local farmers to test the system in the field. These datasets will enable machine learning researchers to build models for predicting fish yield in terms of weight gain, water quality parameters, and feed consumption.

We are excited and grateful to the Meridian Institute for offering us the opportunity through the Lacuna Fund to implement this project that we are passionate about in practical terms. We are hopeful it is going to open a vista of opportunities to local fish farmers as it will throw an avalanche of light into what happens inside the fish pond. This will indeed explain many things to the farmers and improve yields, as well as make available local datasets for the machine learning community.”

Collins Udanor
The University of Nigeria

Locational Offset Correction

The utility of existing crop-cut yield datasets is often compromised due to location inaccuracies arising from the data collection process. To address this, the Locational Offset Correction project will create a method for correcting location inaccuracies using available satellite data. This method will then be used to create a new, clean version of a crop-cut yield dataset for maize crops in East Africa, which will be released along with a tool to allow others to re-create the process using additional datasets. This research will be conducted by a team from Zindi and the Big Data Platform of the CGIAR.

We are looking forward to working with the Lacuna Fund to increase the representation of agriculture datasets in Africa. We will be calling on the amazing data science talent from across Africa and around the world to crowd-source a machine learning solution for correcting location errors, which are a common problem in agriculture datasets. This solution will allow us to correct and publicly release one of the most expansive crop cut yield estimation datasets for maize in Eastern Africa.”

Celina Lee
Zindi

Machine Learning Datasets for Crop Pest and Disease Diagnosis based on Crop Imagery and Spectrometry Data

This project will produce quality open and accessible image and spectrometry datasets from Uganda, Tanzania, Namibia, and Ghana for several crops that contribute to food security in Sub-Saharan Africa, including cassava, maize, beans, bananas, pearl millet, and cocoa. The team -composed of data scientists and researchers from Makerere University, The Nelson-Mandela African Institution of Science and Technology, Namibia University of Science and Technology, and the karaAgro AI Foundation – expect the image and spectral datasets will be used for early disease identification, disease diagnosis, and modelling disease spread, which will ultimately help in breeding resistant crop varieties.

This project is a unique collaboration across four countries in sub-Saharan Africa with the aim of delivering crop imagery and spectrometry datasets for six important food security crops. The datasets are necessary for building machine learning models for early disease diagnosis and will be relevant for not only the AI and machine learning communities but also for the smallholder farmers and agricultural experts.”

Joyce Nakatumba-Nabende
Makerere University