University of Ghana researcher builds AI tool to help farmers spot crop diseases

Pinned summary · as of 20:14

A University of Ghana research assistant, Nana Benyin Eboe Nyamekye Essel-Biney, has created an AI platform named Ndua Analytics. Farmers type or speak what they see on their plants, and the tool suggests which disease or pest may be responsible.

What we know

• The developer of Ndua Analytics is Nana Benyin Eboe Nyamekye Essel-Biney, a research assistant at the University of Ghana and a YARA Research Fellow.
• The tool relies mainly on a farmer's description of symptoms and does not ask for photographs.
• Mr Essel-Biney said he steered away from image-only diagnosis because weak cameras, motion and shadows can distort pictures and make a diagnosis less dependable.
• Testing is under way on maize, cassava and tomatoes. Answers come in English and in a local Ghanaian language, with several voice options.
• Mr Essel-Biney said the approach based on text and speech let him train the model on a fairly small dataset.
• Mr Essel-Biney intends to add rice, onions, ginger and garlic, along with regional data on how disease and pest risks vary, and more Ghanaian languages.
• For now users reach the platform through a web link, and a login system is planned for later versions.
• Prof. Benjamin Lamptey, a meteorologist who is a visiting professor at Leeds, said armyworms and other seasonal pests may show up at unexpected times when conditions favour them, and that researchers are studying whether AI can help forecast their arrival.
• Prof. Lamptey called for researchers, industry and the people who use climate information to work together, and for more investment in both young and experienced researchers.

What's disputed / unconfirmed

• ?
  MyJoyOnline reported that the system gives treatment, prevention and long-term management advice after analysing the symptoms entered; this has not been confirmed by other reports.
• ?
  Chale News reported Mr Essel-Biney as saying the model uses few-shot learning, a method that predicts from limited data, and that the local-language option was chosen to help farmers who struggle with English or technical terms; no other report carries this.
• ?
  Chale News reported that Prof. Lamptey cautioned that technology cannot fix the problem if scientific findings never get to farmers, extension officers and policymakers; this comes from one report only.

Why it matters for Ghana

Many smallholder farmers in Ghana have little access to agricultural extension officers, so a tool that responds to plain descriptions in a local language could help them act on crop problems sooner. Maize, cassava and tomatoes are staple and market crops, so pest and disease losses reach household incomes and food prices. The tool is still being piloted, and the researcher says login features and more crops and languages are still to come.