QUALITY CONTROL RELEVANCE ON ACQUISITION OF LARGE SCALE GEOSPATIAL DATA TO URBAN TERRITORIAL MANAGEMENT

Abstract

Quality control (QC) of geospatial data is relevant to urban territorial management to ensure accurate data for government to make strategic decisions when planning cities. The acquisition and control of geospatial data in the Brazilian government must follow INDE - National Data Spatial Infrastructure - through the Technical Specifications. The cadastral cartography from urban areas in Brasilia was updated and divided into 10 areas. Acquired data includes classes, features, attributes and metadata on 1:1,000 scale. High resolution images and LIDAR data were used to assist the QC process. The first step of the QC was to check positional accuracy. Samples were applied for each class in the mapping block with 4% rate on the feature random selection and all features class had the same level of confidence. Then, three stages were automatically verified: logical consistency, commision and attribute thematic accuracy evaluations. The process also includes the visual interpretation for omission and classification, which involves a certain subjectivity. Everything was executed with QGIS, FME, Erdas Imagine, Postgresql, PostGIS and a plugin specifically developed for that, the DSGTools. The results show that in general, the quantity of errors were low. However, many errors were detected in the elements completeness and thematic accuracy, specially in areas 1, 2, 3, 6 and 9. In the opposite, the logical consistency and positional accuracy presented the lowest quantity of errors, which does not diminish the relevance of these errors, since it compromises the usability of the data.