Sistema de Extração de Malha Viária Rural de Imagens de Satélite usando Redes GAN⋆

Erick P. Barreto; Mateus A. Silveira; Lucas V. R. Alves; Marcelo F. Porto

doi:10.20906/SBAI-SBSE-2023/4018

Erick P. Barreto Colégio Técnico, Universidade Federal de Minas Gerais, MG
Mateus A. Silveira Escola de Engenharia, Universidade Federal de Minas Gerais, MG
Lucas V. R. Alves Colégio Técnico, Universidade Federal de Minas Gerais, MG
Marcelo F. Porto Departamento de Engenharia de Transportes e Geotecnia, Universidade Federal de Minas Gerais, MG

DOI: https://doi.org/10.20906/SBAI-SBSE-2023/4018

Keywords: Road network identification, Satellite images, GAN neural networks, Breadth-first search, Automation in geospatial analysis

Abstract

In this article, we propose a road network identification system that employs a GAN neural network and a breadth-first search-based detection algorithm to identify roads in satellite images. The objective is to evaluate the effectiveness of this technique, aiming for its potential application in automating the road network generation process.