Cameroonian entrepreneurs are using AI to fight crop disease

In Africa, crop pests and disease have been hampering agriculture productivity for decades. Africa farmers lose an estimated 49% of expected total crop yield per annum—the highest in the world – according to the Centre for Agriculture and Biosciences International.

That’s likely to get worse as the impact of climate change worsens according to a study published last year in Science journal. Researchers argue global warming boosts population and insect appetite, and fastens the metabolic rate of insect pests, thereby causing them to consume more. It is estimated that with a 2℃ rise in temperature, insect pests could destroy maize—Africa’s most essential food crop—by as much as 30% more than they do today.

But a team of Cameroonian entrepreneurs has been testing an artificial intelligence based platform to help African farmers tackle crop pests and plant diseases from their source.

Agrix Tech, a Yaoundé-based startup, is planning to roll out its platform across Africa beginning January 2020 when its commercial version will be released. The technology helps detects plant diseases and offers both chemical and physical treatment as well as prevention measures. Founded in August 2018, the Agrix Tech team comprises of a telecoms engineer, certified machine learning engineer, software developer, crop pathologist and two agronomists.

Chiatoh Maryben, an agricultural risk management expert and strategist at Agrix Tech says most farmers in Africa, especially subsistence and small-scale farmers, are not knowledgeable on the types of pests and diseases attacking their crops. “In trying to find a solution, they just use a cocktail of many pesticides. This is very dangerous for humans and the environment.”

But with Agrix Tech, farmers consult the app on their mobile phone, scan a sample of the affected plant and then discover solutions. The app provides both text and voice recognition technology in customized African local languages—as such, even less literate people can use it. Pidgin, Wolof, Fang, French and English are some of the languages which Agrix Tech’s early clients in Senegal, Benin, Morocco and Cameroon have requested.

Adamou Nchange Kouotou, founder and CEO of Agrix Tech says the innovation comes in two forms: a mobile application for organizations which do not have AI teams and as an AI library to help developers add crop disease detection and diagnosis features into their apps. Agrix Tech has already been tested and confirmed by a renowned agro research agency. The innovation’s prototype has a 99% accuracy, according to its founder.

Once the commercial version is rolled out early next year, it will be able to detect multiple diseases in maize, rice, bell pepper, onions, tomato, Irish potato, pepper, mango, lemon, watermelon, cucumber, cabbage, groundnut, citrus, date palm, amongst others.

“We will add features to Agrix Tech as the need arises. We want to mostly cover vegetables because they are more vulnerable to diseases. The most vulnerable are our concern,” Adamou said.

The licence to use Agrix Tech sells at 15 euros a year.