A team of scientists has developed a mobile phone application which uses artificial intelligence (AI) to accurately identify crop diseases in the field.
The app also delivers the latest advice to manage all major diseases and pests that affect root, tuber and banana crops, and helps farmers identify the nearest agricultural extension support for the farmers.
The project which is being implemented by a global network of scientists is part of the Consultative Group for International Agricultural Research’s (CGIAR) research programme on Roots, Tubers and Bananas (RTB).
“As smartphones become more common in rural Africa, they also become handy in agricultural productivity.
“Smallholder farmers or extension officials having basic smartphones with a camera can download the application free of charge, run it up and point the camera at a leaf that has disease indications. They will then get an immediate diagnosis of the disease affecting the plant.” said Dr James Legg, a researcher at the (IITA), in Tanzania, who heads the project alongside Dr David Hughes of Penn State University.
Cassava brown streak and cassava mosaic diseases have for a long time been a threat to food security and income generation of over 30 million farmers in East and Central Africa.
Similarly, the region’s banana production is vulnerable to fungal and bacterial diseases such as the devastating banana bunchy top virus and while late blight which beleaguer potato farmers.
Rural farmers are often incapable of properly identifying these diseases, while researchers, plant health experts and extension officials lack the data to support them, hence the significance of this development.
The current app was developed to help identify cassava diseases, but the team of developers/researchers was awarded Sh10 million in grants as part of CGIAR’s platform for Big Data in Agriculture Inspire Challenge in September, to help them expand the application to other root, tuber and banana crops that are key food sources.
This will in turn boost nutrition and income security for many farmers.
In the application’s initial development, careful fieldwork involving cameras, spectrophotometers and drones at cassava field sites in coastal Tanzania and on farms in western Kenya generated more than 200,000 images of diseased crops to train the system’s AI algorithms.
Using these images, the scientists advanced an AI process that is able to automatically classify five cassava diseases, and by involving tech company, Google, the team was able to develop the smartphone application using TensorFlow, an open-source software library for machine learning across a range of tasks.
The system is currently under field-test in Tanzania.
Penn State University has also developed a mobile spectrophotometer through a small firm called Croptix, whose initial results indicate it can accurately diagnose different viral diseases in the field, even when the plant looks healthy.
“The application similarly uses AI in real time so the farmer can be an active contributor in disease diagnosis and plant health management, hence more yields for smallholder farmers.
“It is similarly groundbreaking because our AI is based on research from scientists at CGIAR and RTB, who are among the world’s best human intelligence on African crops,” said Dr Hughes.
The team has established a working association with Vodafone’s agriculture SMS platform, DigiFarm, which will allow them to link digital diagnostics to largescale text messaging services used by rural farmers.
It will in turn deliver farmer-tailored SMS alerts on crop diseases and pests to 350,000 Kenyan farmers by July 2018.