Agriculture has a greater impact on the environment than any other economic sector in the world. There is a problem of forecasting greenhouse gas emissions depending on various factors of agricultural activity. It is proved that the ability to make forecasts of the amount of greenhouse gas emissions would significantly increase the rational use of agricultural resources. The use of an artificial neural network as a forecasting method in the field of agriculture can allow predicting the environmental impact of greenhouse gases in advance. In addition, this method of data analysis can be an alternative solution of agroecological problems. The article clarifies the features of constructing artificial neural networks for the analysis of dynamic data in the field of agricultural ecology. The approach to solving the problem of forecasting greenhouse gases emissions -an artificial neural network is considered. The problem of forecasting greenhouse gases is considered as a problem of forecasting time series.As an example, an artificial neural network has been designed to predict CO 2 eqemissions from agriculture in the Russian Federation. The data for training the neural network were taken from the Food and Agriculture Organization open database (FAOSTAT).The test results of the developed neural network are presented.The conclusion about the possibility of further appl ication of the development (the model of an artificial neural networkfor predicting СО 2 eq emissionsin agriculture) is given.

This article describes the basic principles of constructing artificial neural networks for pattern recognition on raster images using robotic complexes. The basic parts of convolutional neural networks are considered. Examples of evaluating the accuracy of preliminary testing results developed by the authors of a convolutional neural network model are given.