When commercial-scale crops in First World countries get diseases, lab-equipped experts are typically called in to identify the affliction and advise treatment. Such resources aren't always available to smallhold farmers in developing nations, however, who may lose entire crops without ever knowing what was wrong with them. That's why scientists are now creating software that could be incorporated into an app that identifies crop diseases, based on user-supplied smartphone photos.
The system is being developed by researchers at Penn State University and the Swiss Federal Institute of Technology (EPFL).
To start, the scientists built a model of the system by linking a cluster of computers together to form a neural network. They then fed that model a database of over 53,000 photos of diseased and healthy plants – 14 crop types and 26 diseases were represented.
Utilizing a deep learning approach, they trained the model to look for patterns in all that visual data. Ultimately, the system was able to identify both crops and diseases – from photos – with an accuracy rate of up to 99.35 percent.
Although the algorithms required to initially run the model required significant processing power, the same functions could reportedly now be programmed into an app using just a few lines of code. It's been likened to the manner in which smartphone cameras are currently able to identify faces.
"Given the expectation that more than 5 billion smartphones will be in use around the world by 2020 – almost a billion of them in Africa – we do believe that the approach represents a viable additional method to help prevent yield loss," says Penn State's Prof. David Hughes, co-author of the study. "With the ever-improving number and quality of sensors on mobile devices, we consider it likely that highly accurate diagnoses via the smartphone are only a question of time."
Source: Penn State
I wonder if this work can be combined with a drone based scanning sensor base in development which can detect diseased areas of the crop. Adding mineral and water remote sensors would make this a complete system for the farmer to maintain his crops at an optimal level.