INFORMATION AND MEASUREMENT SYSTEM FOR NON-DESTRUCTIVE HYPERSPECTRAL CONTROL AND DIAGNOSTICS OF THE CONDITION OF CULTIVATED PLANTS

An information and measurement system for remote and proximal sensing of agricultural crops is presented. An example of its application in an intensive apple orchard is shown. The method of hyperspectral imaging in the range from 350 to 1000 nm of defects caused by phyto-diseases and pests was used. The sensing is carried out using a developed system installed on an unmanned multirotor aircraft. Diagnostics of the condition of plants is carried out in two stages. At the first stage, remote sensing is carried out from the height of 10... 15 m, aimed at determining the coordinates of problematic areas of the garden or individual trees potentially affected by phyto-diseases. At the second stage, proximal sensing is performed from a height of up to 1 m above the object of control, aimed at determining the type of disease. The description of the technical, informational and methodological support of the system is given. Examples of mathematical processing of hyperspectral images of apple fruits of three pomological varieties, the use of PCA analysis to determine the wavelengths used as independent variables in the construction of discriminant models for the classification of apple plant tissues with an accuracy of up to 90 % are shown.