Siyrak fotometrik ma'lumotlar asosida shahar landshaftining semantik 3D modellarini yaratish va urbanistik tahlil qilish algoritmlari

Abstract

The rapid expansion of urban environments demands efficient computational methods for spatial analysis and architectural documentation. This study introduces an optimized algorithmic framework designed to generate high-fidelity, semantic three-dimensional urban models relying exclusively on sparse photometric data. By leveraging advanced neural radiance fields combined with semantic segmentation modules, the proposed methodology overcomes the traditional limitations of multi-view stereo techniques, which strictly require dense image overlap. The approach facilitates the automated identification and categorization of structural elements within the generated spatial geometry. Analytical evaluations confirm that integrating learned shape priors significantly stabilizes the reconstruction process under data-deficient conditions. The conceptual findings offer a robust digital platform for urban planners, enabling accelerated spatial analysis, infrastructural monitoring, and the development of intelligent city management systems without the necessity for exhaustive preliminary data collection.