Photo-powered app could help farmers diagnose crop diseases themselves

A new app that uses a simple photograph of a leaf to provide near-instant automatic diagnosis of diseased crops could prove a major agricultural breakthrough for farmers and producers alike.

The idea of the Crop Disease Detector app is to assist arable farmers to improve their yield by giving them quicker and more accurate data on which to base their disease control strategies and therefore stop their harvests being diminished by infection.

It has been developed by staff at Manchester Metropolitan University.

40% of worldwide crops lost to disease

Professor Liangxiu Han, Professor of Computer Science at the University’s Department of Computing and Mathematics, said: “Crop plants can be affected by various diseases and it is estimated that almost 40 per cent of worldwide crops are lost to diseases, with the potential to cause devastating economical, social and ecological losses.

“Currently the disease diagnosis mainly relies on human inspection by surveyor.

“However, this method tends to be time-consuming, costly and inaccurate.

“For example, the surveyors have to be regularly trained to maintain quality, which is costly.

“The level of detail is variable and inevitably errors often arise from the subjectivity of individuals and the tedious nature of the task, which could lead to bias and misdiagnosis.  

“In addition, with large areas or fields to be inspected, the scarcity of the trained surveyors makes the monitoring of disease a challenging task.”

As simple as taking a photo

With the Crop Disease Detector app, all the smartphone user has to do is take a photograph of a leaf and the app connects to an innovative cloud-based processing system that allows it to automatically identify abnormalities on the leaf surface.

It takes less than a minute to provide an on-screen diagnosis and an accompanying description of the disease, a severity indicator and information on typical treatment techniques.

The app can depict the leaf images as markers on a map so farmers and producers can pinpoint disease-hit and healthy tracts of land.

Other apps in this field provide library images of diseased leaves against which the user still has to make a difficult visual comparison with the leaf in front of them.

Expertise in big data and image pattern recognition

Prof Han said: “Here at Manchester Metropolitan we have big data expertise and image pattern recognition expertise and we have combined them to develop an automated approach for the accurate and timely diagnosis of crop disease to address the challenges.

“The advantages of our approach are to give farmers and growers more accurate information and to allow a non-expert to perform almost immediate diagnosis.

“Our approach is designed to provide real-time info for crop disease detection, ensuring food security and significantly reducing the cost of disease monitoring.”

Programmed to identity three diseases

Currently the app – which is at the working prototype stage - has been programmed to identify three types of foliar disease that affect predominately wheat.

These are septoria, yellow rust - also known as stripe rust - and leaf rust, which are caused by different pathogens.

Prof Han said: “We chose them because they are the most serious and most common types of disease.

“But the app could very easily be trained to detect a wider range of diseases and in other cereals, plants and trees.”

In collaboration with Fera Science – a public-private joint venture partnership between Capita and Department for Environment, Food and Rural Affairs that provides translational science for the food, plant protection and related industries – and the Institution of Remote Sensing and Digital Earth at Chinese Academy of Sciences, the University is currently looking for opportunities for further validation with more data, and potential commercialisation.

The app development project was funded by an Agri-Tech in China: Newton Network+ Proof of Concept Award administered by consortium leader Rothamsted Research, which is delivering Newton Fund funding with support from the Science and Technology Facilities Council.