Russia Finds a Way to Detect Plant Diseases Early Using AI
The approach can be used for remote monitoring of agricultural land.
Researchers from the Advanced Engineering School “Digital Engineering” at Peter the Great St. Petersburg Polytechnic University, together with scientists from the All-Russian Institute of Plant Protection, have presented a method for diagnosing crop diseases at a pre-symptomatic stage.
The approach is based on artificial intelligence applied to hyperspectral imaging data. This technology records how light is reflected across dozens or even hundreds of narrow spectral bands, making it possible to identify early signs of plant disease before any visible changes appear.
In the future, the method could be used in remote monitoring systems for agricultural land, including drones and satellites, to detect diseases and stress conditions in crops at an early stage.
Proven in Experiments
The experiments were conducted on wheat grown under laboratory conditions, with the setup designed to closely resemble real field environments. Using hyperspectral imaging, researchers collected 864 images.
Taking into account the variability of field conditions, the team developed a distortion-resistant algorithm for sequential preprocessing of hyperspectral data. Artificial intelligence and machine learning were used to achieve stable results.
According to lead researcher Alexander Fedotov, the success of the system depends directly on proper data preprocessing, which allows machine-learning algorithms to reliably distinguish between healthy and infected plants despite various sources of interference.
