Simultaneous graphic generalization of vector data sets

Lars Harrie, T Sarjakoski

Research output: Contribution to journalArticlepeer-review

44 Citations (SciVal)


Manual cartographic generalization is a simultaneous process. However, most automatic approaches so far have been sequential; generalization operators are applied one at a time in a certain order. This has been the case both for model generalization (generalization of the conceptual model) and graphic generalization. Our research seeks to demonstrate that the graphic part of cartographic generalization can be formulated as an optimization problem and accordingly be solved in a single step. This paper deals with several issues regarding this optimization approach. Firstly, a set of appropriate analytical constraints for the generalization process is given, as well as rules for when to apply these constraints. In our approach, we are limited to formulating these constraints on point locations. Secondly, least-squares adjustment is proposed to find the optimal solution according to the constraints. Finally, the conjugate-gradients method is recommended for solving the normal equations. A prototype system for simultaneous graphic generalization has been implemented in C++, which communicates with a commercial map production system. Results from three tests of the prototype system are included in the paper.
Original languageEnglish
Pages (from-to)233-261
Issue number3
Publication statusPublished - 2002

Subject classification (UKÄ)

  • Civil Engineering


  • conjugate-gradients method
  • Delaunay triangulation
  • least-squares adjustment
  • map generalization
  • graphic generalization


Dive into the research topics of 'Simultaneous graphic generalization of vector data sets'. Together they form a unique fingerprint.

Cite this