Meet the Winners: Eyes on the Ground Challenges
Join Zindi, CGIAR and ACRE Africa to discuss advances in crop damage recognition in Africa, and meet the winners of the recent CGIAR-funded challenges on Zindi.
7:00 - 7:05 pm | Introductions & opening remarks - Paul Kennedy
7:05 - 7:10 pm | Introduction to CGIAR - Jawoo Koo
7:10 - 7:15 pm | Introduction to ACRE Africa - Diana Machogu
7:15 - 7:30 pm | Competition 1 winner - Mohamed Abdelrazik
7:30 - 7:55 pm | Competition 2 winner - Damola Oriola
7:55 - 8:00 pm | Wrap up - Paul Kennedy
About the challenges:
The "Eyes on the Ground'' project is a collaboration between ACRE Africa, the Kenya Agricultural & Livestock Research Organization (KALRO), the International Food Policy Research Institute (IFPRI), and the Lacuna Fund, to create a large machine learning (ML) dataset that provides a close-up view of smallholder farmer's fields, with the aim of developing a Picture Based Insurance framework.
In order to help farmers across Africa manage agricultural risk, ACRE Africa uses image data to settle insurance claims and carry out loss assessment. ACRE has partnered with KALRO to review smartphone pictures of insured crops sent in by farmers to verify whether a farmer’s crops are damaged and to provide agricultural advisories. These advisories depend on whether a crop is damaged, and what the cause is of that damage, for instance whether the damage was related to weather, pests and diseases, or man-made factors such as fire, to evaluate an insurance claim and determine appropriate compensation.
Evaluating images for thousands of insured smallholder farmers to verify insurance claims and provide personalized agricultural advisories is time-consuming, slowing down claims settlement and increasing costs of the advisory service. ACRE Africa and KALRO are therefore looking at artificial intelligence to automate image processing, building on data that ACRE Africa, KALRO and IFPRI produced with support from the Lacuna Fund.
CGIAR Crop Damage Classification Challenge - https://bit.ly/3QYKp3j
The objective of this challenge is to create a machine-learning algorithm to classify crops into categories: Good growth (G), Drought (DR), Nutrient Deficient (ND), Weed (WD), and Other (including pest, disease or wind damage). The data for this challenge is a collection of smartphone images of crops.
We invite you to build a model to classify damage type across multiple seasons. By knowing what type of damage a crop experiences, images can be fed into a model to indicate whether a crop was damaged, and needs to be evaluated for insurance payouts. KALRO and ACRE Africa’s personalized advisories for farmers will also depend on the classification of a farmer’s crop into these five categories.
CGIAR Eyes on the Ground Challenge - https://bit.ly/43Fb3Cr
Since most claims are related to drought, this challenge will ask participants to predict drought damage from smartphone images of crops taken in the past. The Eyes-on-the-Ground project has already successfully trained models to predict drought damage in the first two seasons, but those models did not transfer well into the third season on which data are available.
We therefore invite you to improve on the existing solutions, and improve the accuracy at which your model predicts drought damage across multiple seasons. The extent of drought damage can take on values of zero (in case there is no drought damage, and no insurance payout should be made), or a positive value (in case there is drought damage, and an insurance payout proportional to the amount of damage should be made).
We thank CGIAR Research Initiative on Digital Innovation for technical and financial supports that made this challenge possible.
Watch video Meet the Winners: Eyes on the Ground Challenges online without registration, duration hours minute second in high quality. This video was added by user Zindi Africa 18 March 2024, don't forget to share it with your friends and acquaintances, it has been viewed on our site 34 once and liked it 1 people.