Method of Verifying Terrain Based on Analysis of Visual Features

The article presents an approach to terrain verification based on the analysis of visual features in distributed information systems. The proposed approach is based on formalising the process of spatial matching of reference and current visual scenes, followed by a quantitative assessment of their correspondence. A step-by-step mathematical model has been developed that takes into account the presence of errors, incomplete data, and variable conditions for the formation and transmission of visual information. The paper introduces a terrain verification criterion that allows making an informed decision about the correspondence of a scene to a given area of space. The effectiveness of the proposed approach is confirmed by the results of simulation modelling, which demonstrate the dependence of verification accuracy on the noise level, the number of visual correspondences, and the threshold parameters of the model. The purpose develop a method for verifying terrain based on the analysis of visual features and a corresponding mathematical model that provides a quantitative assessment of the scene's correspondence to a given terrain area. Results. To evaluate the effectiveness of the proposed terrain verification method, a series of simulation experiments were conducted in the Python environment. The experiments investigated the influence of the main parameters of the mathematical model on the accuracy of the decision regarding the correspondence of the current visual scene to a given terrain area. In particular, the influence of the noise level in the input data, the number of visual matches, and the choice of the verification criterion threshold value were analysed. Conclusions. As a result of the study, a method for verifying terrain based on the analysis of visual features was developed and experimentally tested, which provides a formalised decision-making process regarding the correspondence of a visual scene to a given area of space. Simulation modelling confirmed the stability of the proposed mathematical model to noise, incomplete data and changes in basic parameters, which indicates the feasibility of applying the method as part of information technologies for processing visual data in distributed systems.