This study is devoted to the automatization of the image masks’ construction of large sized agricultural objects in precision farming tasks for training neural network methods for crop’s nitrogen status analysis using georeferenced images. The scientific direction is extremely relevant because it allows to automate and replace the manual process of data labeling, significantly reducing the cost of preparing training samples. In the paper, four new synthetic data generation methods are proposed for training neural networks aimed at UAV image segmentation by the level of crop nitrogen supply on an agricultural field. In particular, the paper gives a description of synthetic data generation algorithms based on nitrogen covering with lines, parabolas, and areas. Experiments were carried out to test and evaluate the quality of these algorithms using eight modern image segmentation methods: two classical machine learning methods (Random Forest and XGBoost), four convolutional neural network methods based on U-Net architecture, and two transformers (TransUnet and UnetR). The results showed that two algorithms based on areas gave the best accuracy for convolutional neural networks and transformers — 98–100 %. Classical machine learning methods showed very low values for all quality metrics — 27–44 %.

Еach model for forecasting agrometeorological risks based on the analysis of one-dimensional time series is effective for a certain range of initial information. In addition, the values of the initial observations can differ significantly for each specific case, respectively, the widespread use of one method for the analysis of arbitrary information can lead to significant inaccuracies. Thus, the problem of choosing a forecasting method for the initial set of agrometeorological data arises. In this regard, a universal adaptive probabilistic-statistical approach to predicting agrometeorological risks is proposed, which makes it possible to solve the problem of choosing a model. The article presents the results of the first stage of research carried out with the financial support of the Ministry of Education and Science of the Russian Federation a brief overview of the current state of research in this direction is presented, theoretical foundations for predicting agrometeorological risks for a possible onset of drought and frost have been developed, including the task of generating initial information, a description of basic forecasting models, and also a direct description of the proposed approach with a presentation of the general structure of an intelligent system, on the basis of which the corresponding algorithm can be developed and automated as directions for further work.

EXPERIENCE WITH THE USE OF MATHEMATICAL STATISTICS METHODS FOR ASSESSMENT OF AGRICULTURAL PLANTS STATUS V.M. Bure, A.F. Petrushin, E.P. Mitrofanov, O.A. Mitrofanova, V. Denisov