Geo-spatial analysis of urbanization and environmental changes with deep neural networks: Insights from a three-decade study in Kerch peninsula

Abstract

This study presents a comprehensive analysis of land use and land cover (LULC) changes on the Kerch Peninsula over the last thirty years, utilizing advanced satellite data and spatial modeling techniques. The research used Landsat 5, 7 and 8 satellite images to capture the intricate dynamics of LULC changes from 1990 to 2020. A quantitative approach was adopted, involving the use of convolutional neural networks (CNN) for enhanced classification accuracy. This methodology allowed for a detailed and precise identification of various LULC classes, revealing significant trends and transformations in the region's landscape. The spatial modeling incorporated in this study allowed exploration of both large-scale patterns and localized changes, providing insights into the drivers and consequences of LULC dynamics. The statistical analysis revealed a notable increase in urbanized areas, coupled with a decline in natural ecosystems such as forests and wetlands. These changes reflect the impact of sustained urban growth and agricultural expansion, underscoring the need for informed land management and conservation strategies. The study findings contribute to understanding urbanization processes and their ecological implications, providing valuable guidance for sustainable regional planning and environmental protection.