Classification of elements-at-risk using geospatial data for flood vulnerability assessment

Abstract

Floods have been causing substantial monetary damage to a nation's economy. There are no losses when the flooded area does not have any elements on it. These elements include a human system, built environment system and natural system that are at risk of flooding in a given area. The risk is a combination of hazard and vulnerability. Geospatial technologies are widely been used in hazard assessment. This study aims to classify the elements-at-risk on the basis of using high-resolution remote sensing data in order to extract elements-at-risk for flood vulnerability assessment. First, an object-based image analysis performed on the satellite image for classification of Land Use Land Cover (LULC) map. Characterization of buildings features is based on green area, impervious area, vacant area and buildings area by calculating the percentage of LULC map for each regular grid using zonal statistical analysis. The results show that the double storey confined brick masonry building produced a highest accuracy of 48%, followed by Single storey reinforced frame masonry building with 23% accuracy, single storey confined brick masonry building has 16% accuracy, the accuracy of the double storey or higher concrete reinforced frame masonry building is 10%. Wood/zinc/cement-board building produced the lowest accuracy of 1%. The overall accuracy for classification structure element is 50%. Further analysis should be carried out in order to refine the accuracy in order to carry out details assessment of flood vulnerability.