Remote sensing monitoring of urbanized areas

  • Uhl A.V., Dr.Tech in Engineering, Professor Lesya Ukrainka Volyn National University
  • Melnyk O.V., PhD in Engineering, associate professor Lesya Ukrainka Volyn National University
  • Melnyk Y.A., PhD in Engineering, associate professor Lutsk National Technical University
  • Melniychuk M.M., PhD student Lesya Ukrainka Volyn National University
Keywords: urban development, urban areas, remote sensing, image classification, CART algorithm, Google Earth Engine.

Abstract

Urban growth directly affects urban development processes at both the city and district level, but on the other hand, land and housing policies, current built-up area, available land for development, changes in the number of households, population, economy also affect growth etc. Remote sensing and GIS techniques can be used to analyze and model urban growth. This study was conducted to identify and analyze the growth of the city of Lutsk. The work uses several statistical methods to recognize and analyze this regularity, which shows the dynamics of the city's development in the period from 2016 to 2022. In addition to the obtained results, this study also demonstrates the potential of remote sensing, GIS methods, and the capabilities of Google Earth Engine cloud services to analyze urban growth.

To study the expansion of the urban study area, the Dynamic World dataset, which is near-real-time (NRT) with a resolution of 10 m, was used as a basis, which includes the distribution and information of nine classes of land use structure and land cover (Land Use/Land Cover (LULC)).

The Sentinel-2 MSI image collection is available from 06/27/2015 to the present. Sentinel-2 satellites are repeated every 2 to 5 days depending on the latitude. Dynamic world forecasts are generated for Sentinel-2 L1C images with maximum cloud cover less than 35%. Forecasts are masked to remove clouds and cloud shadows using a combination of the S2 Cloud Probability, Cloud Displacement Index, and Directional Distance Transform algorithms. In total, 615 Sentinel-2 images were used in the work for the period from March 25, 2016 to November 9, 2022.

In this study, the method of controlled classification according to the CART (Classification and Regression Tree) algorithm was used.

The results of this study demonstrated the following facts:

  • the city demonstrates stable growth over different periods of time,
    • the maximum expansion of the built-up area was observed in the southern and southeastern directions and fragmentary construction in other directions
    • research has established that the expansion of the outskirts is dominant in all selected periods and occurs in part of the suburban areas of the city.

Due to the limited areas of the city and free lands, it is necessary to carry out rational control and management of urban growth and dynamics.

The proposed technique can be used to analyze not only urban areas, but also other types of land cover and land use.

Published
2023-01-12