Pioneering AI-driven spectral imaging: NOVATERRA breakthrough peacock spot detection system for olive groves

NOVATERRA focuses on reducing diseases in olive groves, in particular the peacock spot, also known as Cycloconium, a fungal disease caused by the fungus Cycloconium oleaginum, which gets its name from the distinctive circular or oval spots it creates on the leaves, which resemble the eye feathers of a peacock. 

Peacock spot is most common in humid conditions and can lead to defoliation, reduced growth, and lower fruit quality. To control the disease, fungicides are applied, but good orchard sanitation practices (removing fallen leaves and pruning infected branches) can also help reduce the spread of the disease.

To reduce pesticide usage and help keep olive groves free from peacock spot, NOVATERRA Work Package 3 has been developing a peacock spot detection system combining spectral imaging and Artificial Intelligence (AI). Plant disease detection using spectral imaging is a technique that uses advanced imaging technology to detect plant diseases, while spectral imaging involves capturing images of plants using specialised cameras that can detect specific wavelengths of light, enabling the detection of changes in the spectral signature of plants indicative of disease. 

To develop its disease detection system, NOVATERRA determined the most important wavelengths for detecting peacock spot in the visible and Near Infrared Region (700nm, 725nm, and 825nm) through lab experiments. Moreover, it tested several different AI models to find the best-performing one, which identified the disease in field images with an accuracy of ≈90%.

Currently, a simple-to-use, user-friendly plug-n-play solution that requires the end user to only plug in the spectral camera in a laptop to detect the disease is being tested in the commercial olive orchards of Myrolion, a member of NOVATERRA consortium, with promising results.

By Ioannis Malounas, Agricultural University of Athens (AUA)