Data to Cartography New MDE-Based Approach for Urban Satellite Image Classification

Abstract

Monitoring of the earth’s surface has been significantly improved thanks to optical remote sensing by satellites such as SPOT, Landsat and Sentinel-2, which produce vast datasets. The processing of this data, often referred to as Big Data, is essential for decision-making, requiring the application of advanced algorithms to analyze changes in land cover. In the age of artificial intelligence, supervised machine learning algorithms are widely used, although their application in urban contexts remains complex. Researchers have to evaluate and tune various algorithms according to assumptions and experiments, which requires time and resources. This paper presents a meta-modeling approach for urban satellite image classification, using model-driven engineering techniques. The aim is to provide urban planners with standardized solutions for geospatial processing, promoting reusability and interoperability. Formalization includes the creation of a knowledge base and the modeling of processing chains to analyze land use.