Blackcurrant Plantations Monitoring Using Drones

The work is dedicated to the study of drone use in horticulture, focusing on an example of blackcurrant cultivation. The research aims to use drones to monitor vegetation in plantations and to maintain the technological environment of plants, using traditional agrotechnical field care methods. The concept offers mapping and instance segmentation followed by multi-label classification operations, taking into account the specifics of blackcurrant plantations. The mapping operation creates blackcurrant plantation maps from images taken by drones at low altitudes. This ensures the acquisition of high-quality maps of large areas with the help of simple image photography cameras. Instance segmentation is intended for extracting singular leaf instances from mapped images, which are analyzed using classification methods to detect blackcurrant diseases, pest spread, nutrient and moisture deficiencies, and other plant vegetation-related parameters. Classification employs machine learning techniques and is specific to the cultivation of a particular plants - blackcurrants. The proposed technology, with appropriate adjustments, can also be applied to the vegetation monitoring of other horticultural plants.