Announcing the Winners of Radiant Earth’s Competition for Crop Detection in Africa
By Hamed Alemohammad, Chief Data Scientist, Radiant Earth Foundation
Five Data Scientists emerged as winners of Radiant Earth Foundation’s competition, in partnership with Zindi Africa, to create a machine learning model that classifies farm fields in Kenya by crop type using time series of Sentinel-2 satellite imagery collected during the growing season.
Earth observations provide critical data for agricultural monitoring at scale, and machine learning (ML) techniques are best suited to learn from these data. Yet, building agricultural ML models poses a problem in Africa due to limited training data, as well as add-on hurdles created by the relatively small size of the farms. These difficulties prompted Radiant Earth to design a competition to crowdsource data science skills globally for the best crop detection model.
Ground reference data for the competition was collected by the PlantVillage team, a research and development unit of Penn State University, which empowers smallholder farmers with cheap and affordable technology. PlantVillage’s work democratizes access to knowledge that can help farmers grow more food.
Crowdsourcing Data Solutions
Organized as part of the Computer Vision for Agriculture (CV4A) Workshop at the ICRL conference on April 26, the data challenge ran from February 2 — March 28, 2020. The competition performed on Zindi, a data science competition platform dedicated to solving Africa’s most pressing problems by bringing together a community of data scientists who collaborate and compete to come up with the best possible solutions.
A total of 440 data scientists across the world participated in building a machine learning model for classifying crop types in farms across Western Kenya using training data hosted on Radiant MLHub. The training data contained crop types for a total of more than 4,000 fields (3,286 in the training and 1,402 in the testing datasets). Seven different crop classes were included in the dataset, including: 1) Maize, 2) Cassava, 3) Common Bean, 4) Maize & Common Bean (intercropping), 5) Maize & Cassava (intercropping), 6) Maize & Soybean (intercropping), 7) Cassava & Common Bean (intercropping). Two major challenges with this dataset were class imbalance and the intercropping classes that are a common pattern in smallholder farms in Africa.
Participants were provided with temporal observations of 12 bands from Sentinel-2 L2A product (ultra-blue, blue, green, red, near-infrared (VNIR), and short wave infrared (SWIR) spectra), as well as cloud probabilities during the growing season.
The Five Winners
The first place is awarded to Karim Amer from Egypt. His award includes a cash prize of USD 1,500 and an invitation to present at the CV4A workshop at the ICRL conference.
The first place award for an African citizen currently residing on the continent is presented to Femi Sotonwa from Nigeria. He wins USD 1,000 in cash, a 1-year subscription to the Descartes Labs Platform, and an invitation to present at the CV4A workshop at the ICRL conference.
The Naver prize for the first place female-identified African citizen currently residing in Africa is awarded to Ansem Chaieb from Tunisia. She receives a USD 1,000 plus an invitation to present at the CV4A workshop at the ICRL conference.
The 2nd place overall award is presented jointly to Mohamed Jedidi (Tunisia) and Lawrence Moruye (Kenya). They receive a USD 1,000 cash and an invitation to participate in the CV4A workshop at the ICRL conference.
The 3rd place overall award is presented to Olayinka Fadahunsi and Michael Okeyode from Nigeria. They receive USD 500 in cash and an invitation to participate in the CV4A workshop at the ICRL conference.
The prizes for this competition were sponsored by Microsoft AI for Earth, Descartes Labs, Naver Labs, Facebook, Partnership on AI, and Google. We are grateful for their support.
Thank you to everyone who participated in our competition and helped to make it a success! Please join us on April 26 at the virtual workshop on Computer Vision for Agriculture virtual workshop (CV4A), where winners will present their solutions.
