Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back A Technique for Obtaining an Optimal Neural Network Training Sample for Evaluating the Thermophysical Parameters of Materials and Media During Remote Monitoring YU.YU. GROMOV, I.N. ISHCHUK, A.A. DOLGOV, V.V. RODIONOV The article presents a method for obtaining an optimal training sample of an artificial neural network of direct propagation with one hidden layer from the Neural Networks Start package of the MatLAB modeling environment based on model infrared images to solve the problem of estimating the thermophysical parameters of the Earth's surface and objects located on it when conducting remote monitoring in the infrared wavelength range. Keywords: remote monitoring, infrared, thermophysical parameters, artificial neural network, optimal training sample. DOI: 10.25791/pribor.5.2025.1584 Pp. 38-46.