Project Details
Title:
Husker AI (Agriculture)
Details:
Every year, Ghana's 800,000 cocoa farmers crack open millions of cocoa pods to extract the beans inside. What remains, the outer shell called the cocoa pod husk, is left on the farm floor to rot or burnt as waste. Ghana generates an estimated 858,720 tonnes of this material every year, making it one of the largest agricultural waste streams in the country.
This waste is costly in three ways. First, rotting husks on farm floors spread black pod disease, Ghana's leading cause of cocoa crop loss, destroying 20 to 30% of yields annually. Second, burning them releases nearly 5 million tonnes of CO₂-equivalent emissions from Ghana each year. Third, and most directly, this material has documented nutritional value as livestock feed, yet none of it reaches animals.
At the same time, livestock farmers in Ghana pay 60 to 70% of their total production costs on feed alone, making local meat production expensive and uncompetitive.
The gap between these two problems is an information failure. Farmers have no tool to measure their husks, assess their quality, or find buyers. Feed manufacturers have no reliable way to source CPH. Husker AI closes that gap.
Objective:Measure what farmers have. Give farmers a way to point their phone camera at a pile of cocoa pod husks and instantly know how many they have and how much they weigh, using AI-powered object detection. 2. Assess quality automatically. Classify each batch of husks into one of three quality grades (A, B, or C) based on visual indicators like colour, surface condition, and freshness, so farmers and buyers know exactly what they are dealing with without any laboratory testing. 3. Tell farmers what their husks are worth and what to do with them. Based on the grade and quantity, recommend the right livestock animal, the correct feed inclusion rate, any processing steps required to make the husks safe for animals, and a fair selling price benchmarked against current Ghanaian feed ingredient market prices. 4. Connect supply to demand. Link cocoa farmers who have husks to aggregators and feed manufacturers who need them, through a built-in marketplace that matches buyers and sellers by location, grade, and quantity. Together, these four goals turn a waste product that currently earns farmers nothing into a measurable, gradeable, and sellable commodity.
Ghana is the world's second largest cocoa producer, and the pod husk waste problem is uniquely concentrated here at scale. The 800,000 smallholder farmers generating this waste earn an average of $1,000 to $1,500 per year, meaning even a modest additional income stream from CPH sales represents a significant livelihood improvement. The solution is designed around Ghanaian realities. The app runs offline because cocoa farms in the Ashanti, Western, and Bono regions have poor internet coverage. Pricing is benchmarked against MOFA's Ghanaian market prices for soya and maize. Feed inclusion recommendations are drawn from research conducted by the Cocoa Research Institute of Ghana. The planned dataset will be collected directly from Ghanaian farms. The project also aligns directly with two active government programmes: the Feed Ghana Programme 2025 to 2028, which targets livestock feed cost reduction, and COCOBOD's ongoing efforts to monetise cocoa byproducts.