Level of detail matters: machine learning based river selection across Polish and French data

Overview & Synopsis

Cartographic generalisation is one of the most important and indispensable elements of the map design process. It is defined as the deliberate and logical reduction of map detail, taking into account the purpose of the map, the desired level of detail (LoD), the perceptual capabilities and limitations of the user, as well as the usage environment (Kraak et al., 2020). One of the first steps in this process is selection, whose goal is to remove detail from higher-resolution datasets until their LoD matches that of lower-resolution datasets. The task of selection is to reduce the map or database content appropriately to the target scale or intended LoD (Stanislawski et al., 2014, Touya et al., 2023). In addition to the most important anthropogenic map features, such as settlements and roads, rivers represent key natural elements. Therefore, it is crucial to develop effective river selection algorithms, especially those that account for multiple LoDs and diverse datasets, in order to create scalable and comprehensive methods and tools applicable across different countries (Le Mao et al,. 2024).