Improving Settlement and Road Network Design for Maps of Small Scales Using Artificial Intelligence and Graph Theory
[Optymalizacja redakcji osadnictwa oraz sieci dróg w skalach przeglądowych z wykorzystaniem sztucznej inteligencji oraz teorii grafów] (2021-2024):
Cartographic generalization is an essential element of map designing. According to the International Cartographic Association definition, generalization concerns the selection and simplification of geographic information relative to the map scale and/or purpose in order to present it at a smaller scale. In this project, we consider selection, which is the first and crucial task to undertake in generalization as first we need to make a decision which objects we will show on the map at a smaller scale.
The development of effective and consistent methodology for generalizing small scale maps has not gained enough attention so far, as most of the conducted research has focused on the acquisition of large-scale maps. Thus we are still a long way from a comprehensive and formalized methodology for small-scale generalization. The presented research aims to close this gap by proposing innovative methods of settlement and road network generalization in maps of small cartographic scales based on artificial intelligence elements (AI), specifically machine learning (ML) and graph theory (GT). Optimal and automatic generalization methods can improve and accelerate the map design process. It can be a response to the growing demand of the information society for current and available data and maps at various scales as it can significantly reduce map design costs.
In this research, the goal is not to achieve a complete reconstruction of the manual cartographer’s work but to automatically achieve the results that would be optimal, acceptable from the cartographic point of view and possibly nearest to the manual map design. The research hypothesis states that the use of machine learning (ML) and graph theory (GT) as a way of the formalization of expert cartographic knowledge taking into account the object features and spatial context can be an effective approach for the multi-aspect and optimal settlement and road network generalization methodology development.
Within the project, two general objectives are defined. First, this research aims to improve and automate settlement selection for representation at small cartographic scales using machine learning (ML) approaches. The second objective is to develop a consistent methodology for the automatic generalization and evaluation of road networks at small cartographic scales. The research scope concerns developing innovative methodology of settlement and road network selection from the source detail level corresponding to 1:250 000 scale to 1:500 000 target scale.
Finansed by: National Science Center Poland (Narodowe Centrum Nauki).
- Izabela Karsznia (Principal Investigator)
- Karolina Sielicka (Investigator)
- Albert Adolf (PhD student)
- Iga Ajdacka (scholarship student)
Promotion of knowledge
- 26.06.2021, I Naukowy Speed-Dating Polska-Szwajcaria, referat: „Optymalizacja redakcji osadnictwa oraz sieci dróg w skalach przeglądowych z wykorzystaniem sztucznej inteligencji oraz teorii grafów” [Improving Settlement and Road Network Design for Maps of Small Scales Using Artificial Intelligence and Graph Theory] I. Karsznia.
- 28.05.2021, Spotkanie przedstawicieli agencji kartograficznych oraz ESRI w ramach User Community for Geospatial Authorities Working Group 1: Map Automation, webinarium pt: „The use of Machine Learning in cartographic generalization. Case studies of settlement and road network automatic selection for maps at small scales” I. Karsznia.