Title

Semantic Infrastructure and Methods to Support Quality Evaluation in

Cartographic Generalization

Author

Frank Brazile

Advisors

Dr. Robert Weibel, Dr. Kurt Brassel

Year published

2000

School

University of Zurich, Geography Department, Spatial Data Handling Division

Issue Summary

“Cartographic Generalization is the process of handling the design trade-offs that occur when creating maps at successively smaller map scales. Cartographic generalization is a very labor intesive process, and requires expert training and experience to master. Cartographers, scientists, and engineers were interested in automating the process, to the extent possible, when computers first were able to draw and print graphics on screens and plotters in the 1950s and 60s. In the beginning of the computer age of maps, researchers developed computer algorithms to automatically draw linear and area features based on information taken from large scale imagery and base maps. These early maps were crude, compared to the output from master-level human cartographers. However, as data was increasingly stored digitially instead of on traditional analog media, it became important to improve automated generalization capabilities. During the 1990s, cartographic systems researchers began working on ways to combine and improve many of the well known, low-level generalization algorithms which previously focused on individual features types, such as line smoothing or area aggregation, to produce a wholistically generalized map. These professional researchers in the cartographic community identified many gaps, and proposed additional algorithms and steps to detect and react to design context. I was fortunate to work with international research teams in the field of wholistic cartographic generalization, which relies heavily on computational geometry, cartographic design, computability theory, software engineering and map production workflow expertise. This dissertation reports on my suggestions to use features found naturally in the map, to help constrain and decompose the optimization problems often found in generalizing base maps. This idea and many others proposed by team members working on the EU-supported ESPRIT research initiative appeared in open published reports, and the Laser-scan Ltd. cartographic software system. Some advanced reports are closed due to the opportunity for commercial exploitation by research partners.“

Frank Brazile,

San Francisco,

September 2005

Link

FBrazile-diss-UniZurich.pdf