Urban Building Footprint Extraction from High Resolution Image for Disaster Damage Assessment

Keywords: Urban disaster, Building footprint, Remote sensing, OBIC
Session Type: Virtual Paper Abstract
Day: Monday
Session Start / End Time: 2/28/2022 03:40 PM (Eastern Time (US & Canada)) - 2/28/2022 05:00 PM (Eastern Time (US & Canada))
Room: Virtual 17

Abstract

Dhaka is prone to earthquakes due to the geotectonic location and rapid urbanization. As urban areas are expanding day by day, studying the environment of the metropolitan regions is becoming essential. Previously, the pixel-based classification method was widely used but in recent times, object-based classification is becoming popular to study urban settings due to the advancement of the process, which offers more accuracy than other methods. In object-based image classification, there are three steps: image segmentation, classification, and accuracy assessment. The classification accuracy depends on image segmentation which has three parameters: scale, shape, and compactness. This study aimed to identify the optimal or near-optimal segmentation approach for extracting building footprint using object-based image classification in urban areas with high-resolution satellite images (GeoEye1), which will help better prepare and respond to urban hazards like an earthquake. Multi-resolution segmentation technique was applied with different values of parameters and near-optimal parameters of scale, shape and compactness were defined. Ruleset classification was applied by using a threshold value of Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), brightness and band values, and then accuracy assessment (Error matrix and Kappa coefficient) was applied. Results show that segmentation parameters directly affect the classification accuracy, mostly scale, and shape. The parameters defined from this study can be used to classify satellite image or LIDAR data for better management of urban disaster (such as earthquake) to identify damages and response in the shortest time after a disaster also in preparation and decision-making in pre-disaster phase.