Application of machine learning methods to analyze changes in water bodies based on satellite images over the past 20 years

In this study, satellite data for 70 typical water bodies in Kazakhstan were collected for the period from 2000 to 2024 using the Landsat 7, 8, 9, and Sentinel-2 missions. The images were processed using existing image segmentation algorithms and deep learning tools. The areas of the water bodies were calculated over different time intervals using artificial intelligence methods. These data and images are intended for monitoring dams and embankments in Kazakhstan. The main goal of data collection is to create a foundation for further analysis and forecasting of the state of water objects.The classification of changes in the area of the embankment was carried out using machine learning methods, where the task was to determine whether the change in area was significant or insignificant. For this purpose, classification algorithms such as logistic regression and other methods were applied, using features such as area and temporal characteristics (year, month, season). The classification results allowed for the identification of two types of changes: significant and insignificant.