Technology for Blackcurrant Plantations Control Using Drones

This article presents a technology-based solution for monitoring blackcurrant vegetation using drones and artificial intelligence. The proposed system, implemented in a blackcurrant farm in Latvia, includes a three-stage process: mapping, identification and segmentation, and classification. Drones capture aerial images of the plantation, which are processed using tools like WebODM and deep learning algorithms to create accurate field maps. Neural networks are employed for identification, instance segmentation and classification of blackcurrant leaves into categories such as healthy, nutrient-deficient, or diseased. The system incorporates several AI model families-YOLO and ResNet -selected based on performance, accuracy, and resource efficiency. The methodology enables high-throughput analysis of large horticultural areas, supporting growers in decision-making by providing precise, visual insights into plant health. The approach demonstrates the viability of integrating drone technology and AI for precision agriculture, particularly in the specialized context of blackcurrant farming. The proposed technology, with appropriate adjustments, can also be applied to the vegetation monitoring of other horticultural crops.