Identification of pervious and impervious surface areas (ISA) using GEographic Object-Based Image Analysis (GEOBIA) and deep learning as input to drive a forecast model of urban growth

• Grant number: 20/09215-3
• Support Opportunities: Regular Research Grants
• Duration: February 01, 2021 - January 31, 2023
• Field of knowledge: Physical Sciences and Mathematics - Geosciences
• Principal Investigator: Cláudia Maria de Almeida
• Grantee: Cláudia Maria de Almeida
• Host Institution: Instituto Nacional de Pesquisas Espaciais (INPE). Ministério da Ciência, Tecnologia e Inovações (Brasil). São José dos Campos , SP, Brazil
• Associated researchers: Camila Souza dos Anjos Lacerda ; Elcio Hideiti Shiguemori ; Gustavo Barbosa Lima da Silva

Abstract

In the two last centuries, particularly in the latest decades, humankind has witnessed a drastic shift of population from rural to urban areas. Despite the small area occupied by urban areas compared with the total habitable land, urban centers are responsible for the highest proportion of environmental impact. Some of the most common modifications caused by urbanization involve land use and land cover change, increased impervious surface cover, altered hydrology, higher temperatures, and elevated air, noise, and visual pollution. In this context, the amount of impervious surface areas (ISA), which is directly related to population growth and urbanization, is a key parameter in determining the impact of urbanization on the environment and a major indicator of environmental quality. Understanding and anticipating the impacts of urbanization requires detailed characterization of the current distribution of ISA within urban settlements, as well as future trends of their expansion. In this way, the goal of this project is twofold: i) to develop a protocol for ISA mapping using cutting edge approaches based on WorldView-3 satellite imagery, Geographic Object-Based Image Analysis (GEOBIA) and artificial intelligence (Deep Learning - DL) applied in a pilot area; and ii) to use this ISA information in order to drive a DL-parameterized spatial dynamic model to simulate and predict urban growth and land use change. The results of this research are meant to provide subsidies to local authorities and decision-makers in general so as to support related actions in the planning and management of cities growth according to the environmental carrying capacity of concerned sites and to their present and envisaged infrastructure availability. All the approaches to be explored in this work are embedded within the domain of the Big Data paradigm and comply with the emerging concept of Smart Cities. (AU)