ABSTRACT
Large-scale spatial data that include both topographic and thematic information are needed to support decision-making and regular activities at municipal level. Among these, urban planning and management require faster updating of municipal spatial databases. The combination of widely-available, wide-coverage, cost-effective Very-High spatial Resolution (VHR) satellite imagery andGeographic Object-Based Image Analysis (GEOBIA) hold promise for this purpose (Ehlers 2007, Hay & Castilla 2008). However, until recently, there remained spectral limitations of the former (Herold et al. 2003) and some shortcomings of the latter (Lang 2008). In a detailed evaluation of classification performance, Herold et al. (2003) have identified spectral limitations of IKONOS imagery for pixel-based classification of urban land cover, indicating that four-band VHR satellite imagery are not well suited to capture in detail the unique spectral characteristics of the urban environment (Herold et al. 2003).
