University of Twente Student Theses
Improvement of Elevation Model Accuracy and Suitability for Hydrodynamic Modelling
Sarkar, Prangya Paramita (2009) Improvement of Elevation Model Accuracy and Suitability for Hydrodynamic Modelling.
PDF
5MB |
Abstract: | Flooding is one of the most destructive and frequent natural disasters affecting many countries of the world. Increasing precipitation due to climate change is increasing the probability of related phenomena such as floods in future. Thus, there is a need of efficient flood risk management. Flood models have a great potential to aid flood risk management. The most important component of flood models is the surface elevation information. The efficiency of prediction about different derivatives of water flow during a flood event depends upon the elevation information. Thus, for flood event analysis and flood risk management, the elevation information should be accurate. Photogrammetrically derived elevation models represent the earth’s surface topography along with the surface features on it. Some of the surface features are impermeable obstacle to water flow (e.g. buildings) and some of them are permeable (e.g. trees). The presence of actually impermeable obstacles in the elevation model can influence the flow of water and thus can make the flood model prediction erroneous. Therefore, those impermeable obstacles should be eliminated from the elevation model and the permeable obstacles should be retained there. From this motivation the present study was undertaken to develop a semi-automatic method to remove the impermeable obstacles and retain the permeable obstacles in the photogrammetrically derived surface model. This research used scanned coloured aerial photo to generate a digital surface model (DSM) and orthophoto of the study area. The method was formulated, firstly, to identify the features and secondly, remove relevant features selectively from the DSM. Only pixel based analysis was taken into consideration to limit the scope of the study. Colour and texture of the landcover features were used to identify the features. Binarisationinterpolation method and neighbourhood analysis were used to remove the relevant features from the DSM. The resultant elevation model is termed as pseudo-DTM. Later on the pseudo-DTMs obtained through different methods were compared with a reference LiDAR DTM to assess their precession and accuracy. The methods were also assessed with respect to different practical scenarios. The present research proposes two different methods for selective removal of the surface features from the DSM. This research is a stepping stone for further exploration in this field using similar technique but with more detailed dataset. |
Item Type: | Essay (Master) |
Faculty: | ITC: Faculty of Geo-information Science and Earth Observation |
Programme: | Geoinformation Science and Earth Observation MSc (75014) |
Link to this item: | https://purl.utwente.nl/essays/92719 |
Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Repository Staff Only: item control page