This work is aimed at investigating the effectiveness of the adapted AKAZE algorithm for processing images of the surface of glaciers, which have a complex structure due to the variety of textures, cracks and variability of lighting conditions. Glacial surfaces pose a challenge to traditional image processing methods, as the qualities of the connecting points can vary significantly depending on the environment. In the study, preliminary image processing was carried out, which made it possible toincrease the accuracy offinding the connecting points necessary for accurate modeling. Using the Python programming language, the AKAZE algorithm was implemented to find and measure connecting points in images. The images were then transferred to the PHOTOMOD 7 software package, where several subsequent processing steps were carried out to create a digital relief model. Afull cycle of processing ofthe same block was also performed using internal algorithms in the PHOTOMOD 7 and Agisoft Metashape software package to compare results and evaluate quality. A comparative analysis of the obtained elevation matrices demonstrated the effectiveness of the AKAZE algorithm in the context of processing complex glacial surfaces. The results of the study highlight its potential in geoinformation tasks and open up new horizons for further research in the field of photogrammetry and monitoring of natural objects.
Olga Nefedova (Fri,) studied this question.