Digital Elevation Model from Unmanned Aerial Vehicle: An Application in Urban Flood Modeling

Reddy, B Sreekarthikeya and Regonda, Satish Kumar (2019) Digital Elevation Model from Unmanned Aerial Vehicle: An Application in Urban Flood Modeling. Masters thesis, Indian institute of technology Hyderabad.

Text
Mtech_Thesis_TD1485_2019.docx
Restricted to Repository staff only until 10 December 2019.
Download (8MB) | Request a copy

Abstract

Migration of people from rural to urban areas is the basis for the urbanization, and it has been happening over the few decades and is on the rise off late. The land use-land cover (LULC) changes due to urbanization consists of increase in the percentage of impervious area, alterations of natural water streams, and modification to existing sewer systems. This results in decreasing the infiltration volumes, consequently increasing the runoff volumes. An another apparent change is decrease in time to achieve peak runoff. These factors together increase the likelihood of flooding in the urban areas. Flow paths and flood inundated areas are key in the context of urban flood modeling either to improve the existing infrastructure to accommodate the floods or to enact the government department concerned in the disaster management. Identification of flows paths and then mapping of inundated areas can be done using different methods. Typically, these methods are data intensive and apply numerical methods. In this study, flow path delineation is done from a Digital Elevation Model (DEM) applying two different flow path identification algorithms. DEMs can be acquired from different types of sources, and one of the sources is images taken from an Unmanned Aerial Vehicle (UAV). Spatial resolution of the DEM plays a key role in the flow path delineation, consequently flood inundation. In this context, one of the objectives of this study is to understand the impact of flight parameters (e.g., flight altitude, camera angle, overlap between the adjacent images) on DEM. The second objective is to develop flow paths using D8 and D-Infinity algorithms. The third objective is to simulate the flow movement and then map flood inundated areas, which requires the application of flow routing. It is the understanding of the impact of flow paths and precipitation on flood inundation. In this research, I used DJI Phantom 4 Pro for the image acquisition, Pix4D software to develop DEMs using a varied combination of flight parameters, and RStudio to perform raster operations and statistical analysis. It was observed that lower camera angles at higher flight altitudes, and higher camera angles at lower flight altitudes, do not provide an accurate representation of the terrain. It was also observed that the combination of 700 camera angle, 70% frontal overlap, and 63% side overlap at any flight altitude provides the best possible representation of the terrain. Flow routing was performed using a finite difference approach to the diffusion wave approximation of the St. Venant equations. MacCormack scheme, which is a second-order accurate, explicit two-step predictor-corrector method, was used to obtain a solution to these equations. The depth of flow, the velocity of flow and discharge flowing out of each pixel in the DEM is determined at each time step, which gives a better understanding of areas that may get flooded during a rainfall of certain intensity.

[error in script]

IITH Creators: