Кафедра комп’ютерних наук

Постійне посилання на фондhttps://repository.lntu.edu.ua/handle/123456789/78

Переглянути

Фільтри

20261

Налаштування

Автор: search.filters.author.Tulashvili, YuriiКвартиль та наукометрична база: search.filters.custom_quartile.Вітчизняні фахові наукові видання

Результати пошуку

Зараз показуємо 1 - 1 з 1
  • Ескіз
    Item type:Наукова стаття,
    Geoinformation web system for assessing the state of carpathian forest ecosystems using remote sensing data
    (2026-04-26) Tulashvili, Yurii; Lukyanchuk, Yuri
    The article addresses the pressing issue of developing modern software tools for automated monitoring of forest ecosystem dynamics in the Ukrainian Carpathians. It is emphasized that traditional forest inventory approaches, based on periodic field expeditions, are insufficiently responsive and economically burdensome, particularly in complex mountainous terrain. The feasibility of integrating Earth remote sensing (ERS) technologies and geographic information systems (GIS) into forestry practices is substantiated. The research aims to develop a web-based information system capable of processing satellite imagery (notably Sentinel-2), calculating spectral vegetation indices (NDVI), and visualizing analytical results through interactive cartographic layers. The paper presents the architecture of the proposed solution, which is grounded in a microservice approach utilizing the FastAPI framework, the PostgreSQL/PostGIS spatial database, and the Leaflet mapping library. The process of designing the logical geodatabase model, ensuring efficient storage of both vector objects (forest compartments and subcompartments) and raster image metadata, is described in detail. The practical implementation of the application’s server-side components is showcased, including mechanisms for asynchronous request handling, JWT authentication, and NDVI calculation algorithms employing the Rasterio and NumPy libraries. The section on specialized calculations provides a comparative performance analysis of spatial SQL queries with and without GiST indexing, confirming a speed enhancement exceeding 100 times. Furthermore, a validation of the vegetation index calculation accuracy was conducted by comparing results with ground-based spectrometry data. It has been established that deploying the developed system reduces the time required for forest area monitoring by a factor of 5–7 compared to traditional ground survey methods, thereby demonstrating the high economic and environmental efficiency of the proposed approach.