Multi-temporal recognition of built-up area and land cover changes using machine learning approach in the Metropolis of Aix-Marseille-Provence in France

• Published in: 2023 Joint Urban Remote Sensing Event (JURSE) pp. 1 - 4
• Main Authors: Gadal, Sebastien, Oukhattar, Mounir, Otobo, Solomon Abody
• Format: Conference Proceeding
• Language: English
• Published: IEEE 17.05.2023
• Subjects: Aix-Marseille-Provence Metropolis; Artificial satellites; Biological system modeling; Earth; Economics; geo-simulation; LCC; Machine learning; Planning; remote sensing; Sociology; Urbanisation
• ISSN: 2642-9535
• DOI: 10.1109/JURSE57346.2023.10144184

Over the last thirty years, the extension of built-up areas has affected all areas of the Aix-Marseille-Provence (AMP) Metropolis. Urban sprawl has been particularly important in generating fragmented urban territories with various forms and spatial patterns. The modelling of the evolution of AMP is fundamental in the context of the implementation of climate plans for the horizon of 2050. It is based on spatial modelling by remote sensing of land cover change (LCC) between 1984 and 2021 using a classification approach that combines spectral transformations and indices applied to Landsat 5 TM and Sentinel 2 MSI.The geo-simulations of the LCC and built-up areas dynamics by 2050 are modelled on Markov chain cellular automata. The future trends of AMP Metropolis are characterised by the evolution of built-up areas, estimated at 4,9% in 1984, 9,6% in 2021 and 11,2% in 2050 mainly to the loss of agricultural lands. The forests and semi-natural environments tend to be mainly more resilient to urban growth. The modelling of spatial dynamics of urbanisation is correlated with those of the evolution of the territorial distribution of populations. Geo-simulation of spatial dynamic changes is one of the decision-making planning tools for better management of the use of AMP territories.