Resumo:
Topological consistency in geospatial data is an essential requirement for territorial management. This work proposes and validates a hybrid methodology for correcting inconsistencies in extensive vector databases, focusing on the production of fiscal thematic cartography in Salvador. The case study started from a universe of more than 1 million CDGV features, composed of homologated and non-homologated sets, which presented 136,142 identified inconsistencies. The methodology was structured in two iterative development flows, based on the PostgreSQL/PostGIS architecture and on the QGIS/DSG Tools tools. The application details the methodological difficulties encountered, such as the introduction of new errors by ST_Snap, and the hybrid solutions developed, such as the use of “negative buffer” in SQL to correct containment errors. The process resulted in an integrated database with 544,746 parcels and 16,182 blocks hierarchically consistent. As a final result, the topological consistency allowed the implementation of Triggers in PostGIS, creating an integrated geospatial database that automates the generation of geocodes and protects itself against future errors. The hybrid methodology, combining SQL, plugins, and manual review, is validated as a robust solution for the treatment of legacy and complex geospatial data.
