Substantiation of a digital system for the identification of infectious diseases and removal of infected potato plants

Relevance . Prevention of the spread of infections at the stages of seed production and breeding of new varieties plays an important role in potato breeding and seed production. Currently, in breeding and seed production, a manual or mechanized method of carrying out variety harvesting is used, which has a low level of automation and implies the presence of an operator culling and removing potato plants. The objectives of the study are the theoretical substantiation of a digital identification system for infectious diseases of potato plants and the development of an algorithm for identifying the coordinates of diseased plants for their subsequent removal. Methods . In the process of developing a system for recognizing infected potato plants and determining the coordinates of their tuber nest, machine vision technologies and deep learning algorithms, as well as a database of images of infected plants, were used. Photogrammetric methods were used to determine the actual size of the object in the frame, as well as to minimize lens distortion and take into account camera angles relative to objects. Results . The necessity of creating a digital identification system for infectious diseases of potato plants for their subsequent effective detection and removal is theoretically substantiated. Factors affecting the accuracy of determining coordinates, such as lens distortion and camera orientation angles, are taken into account. An algorithm for the joint operation of a system for recognizing infected potato plants, a system for determining the coordinates of diseased plants, and an actuating mechanism for extracting plants from the soil is proposed.