The ongoing war in Ukraine has caused extensive damage to infrastructure, agriculture, and the environment, while ground-based assessment remains severely constrained due to security concerns. This paper presents a novel change detection methodology based exclusively on openly available Sentinel-1 and Sentinel-2 satellite data. The key contribution of the proposed approach is the automation of post-conflict change analysis tailored to land-cover type (urban vs. non-urban), achieved through the integration of SAR-based change detection results and optical image classification, combined with the reduction of local classification artifacts using context-aware smoothing. The proposed algorithm enables automatic land-cover type classification and adaptive selection of the appropriate analysis strategy as an outcome of land-cover change assessment using Sentinel-1 and Sentinel-2 imagery. A comparison of the obtained results with the UNOSAT database confirmed the detection of more than 80% of damaged buildings (quality metrics: recall 78.8%, precision 87.5%, F1-score 0.828). The proposed classification method incorporating context-aware smoothing achieves higher built-up area detection accuracy than global land-cover products, outperforming the AlphaEarth platform (0.98 vs. 0.87). The presented approach enables rapid land-cover change analysis and damage detection using openly available satellite data, particularly in conflict-affected regions where direct field measurements are restricted due to security constraints.
Karwowska et al. (2026) studied this question.