Integration of Ikonos data with existing data sources to monitor local impacts of rapid human disturbances and natural hazards

You are here

Context and objectives

The purpose was to compare IKONOS data with various other data sources, for the updating of topographical maps, to produce a vegetation map, which can serve as input in hydrological models, and to map expansion of rural settlements and urbanisation. A further purpose was to present a methodology to integrate multi-source remote sensing data into a consistent time series of land cover maps in order to carry out change detection.

Project outcome

Expected scientific results

IKONOS data have mapping capabilities comparable to the colour aerial photographs used in this study, but have the disadvantage of being much more expensive and less suitable for studies covering large areas. Integration of IKONOS data with high-resolution aerial photographs in a time series analysis allows us to monitor rural settlements and even detect settlement changes linked to the establishment of temporal dwellings. A spatial resolution of 1m, with the fusion of panchromatic and multispectral information of the IKONOS data proved indispensable for the detection of the smaller family units.
Both IKONOS and colour aerial photographs allowed for the mapping of riverine vegetation, which could not be recognised on the lower-resolution imagery. The automatic classification of satellite images clearly is a big advantage compared to the manual interpretation of aerial photographs.
The cost analysis revealed that fluctuations in the quality of remote sensing data can be very high. The potential for interpretation of these data has an direct incidence on the cost (good data will require less fieldwork). For a mapping institution, it is much easier to impose data quality norms (and to control their application) when they order aerial photography. Satellite imagery will be provided in multiple scenes, many having variable acquisition data and quality. The mapping institution has practically no control over the quality, and can be contractually forced to pay for data having actually a low potential for mapping.