Formalization and Data Enrichment for Automated Evaluation of Building Pattern Preservation

Abstract

ABSTRACT:Automated evaluation of generalization output relies to a large extent on that requirements (e.g. specifications, constraints) being formalized in machine-readable formats. Previous studies suggest that the formalization and automated evaluation are relatively easier for legibility constraints (improve the readability of maps) than for preservation constraints (preserving important real-world phenomena). Three major difficulties, i.e., pattern classification and characterization, pattern matching, and constraint formalization, in the automated evaluation of building pattern preservation constraint are analyzed in this paper. A classification of available building patterns is reviewed based on a previous work. In addition, the transition events describing allowed changes for building patterns to preserve during generalization are obtained through the study of existing maps series (from 1:10k to 1:100k). Based on the obtained knowledge on pattern types and acceptable transition events, an approach to automatically match corresponding building patterns at different scales is presented. The methodology proposed is validated by applying it to the interactively generalized data. The result shows promising results and also further improvement in order to apply the method in an overall evaluation to indicate acceptable generalization solutions.

Characterization and Detection of Building Patterns in Cartographic Data: Two Algorithms

“…Also noted that another work on the automated evaluation of building pattern preservation constraint (Zhang et al, 2010) is carrying out based on the detection results reported in this paper.…”

Social Welfare to Assess the Global Legibility of a Generalized Map

“…Experiments show that such an approach allows evaluating individually every element of the user's needs, the cartographic rules and even the respect of the initial data [4]. Nevertheless, difficulties remain in the evaluation of characters of complex group of objects like building alignments [9]. Although it allows a fine description of the legibility of a generalized map, the previous approach only provides the building blocks for a global evaluation of the map.…”

Pattern Classification Approaches to Matching Building Polygons at Multiple Scales

“…Better rules are required to replace Rule 2 in order to distinguish incorrect correspondence and potential n-to-m relationships. A prior matching of building groups as described in Zhang et al (2010) may be helpful.…”