Show simple item record

dc.contributor.advisorMaidment, David R.
dc.contributor.advisorPassalacqua, Paola
dc.creatorSangireddy, Harishen
dc.date.accessioned2011-07-07T19:49:56Zen
dc.date.available2011-07-07T19:49:56Zen
dc.date.issued2011-05en
dc.date.submittedMay 2011en
dc.identifier.urihttp://hdl.handle.net/2152/ETD-UT-2011-05-3151en
dc.descriptiontexten
dc.description.abstractLight Detection and Ranging (Lidar) is a remote sensing technique that provides high resolution range measurements between the laser scanner and Earth’s topography. These range measurements are mapped as 3D point cloud with high accuracy (< 0.1 meters). Depending on the geometry of the illuminated surfaces on earth one or more backscattered echoes are recorded for every pulse emitted by the laser scanner. Lidar has the advantage of being able to create elevation surfaces in 3D, while also having information about the intensity of the returned pulse at each point, thus it can be treated as a spatial and as a spectral data system. The 3D elevation attributes of Lidar data are used in this study to identify possible water surface points quickly and efficiently. The approach incorporates the use of Laplacian curvature computed via wavelets where the wavelets are the first and second order derivatives of a Gaussian kernel. In computer science, a kd-tree is a space-partitioning data structure used for organizing points in a k dimensional space. The 3D point cloud is segmented by using a kd-tree and following this segmentation the neighborhood of each point is identified and Laplacian curvature is computed at each point record. A combination of positive curvature values and elevation measures is used to determine the threshold for identifying possible water surface points in the point cloud. The efficiency and accurate localization of the extracted water surface points are demonstrated by using the Lidar data for Williamson County in Texas. Six different test sites are identified and the results are compared against high resolution imagery. The resulting point features mapped accurately on streams and other water surfaces in the test sites. The combination of curvature and elevation filtering allowed the procedure to omit roads and bridges in the test sites and only identify points that belonged to streams, small ponds and floodplains. This procedure shows the capability of Lidar data for water surface mapping thus providing valuable datasets for a number of applications in geomorphology, hydrology and hydraulics.en
dc.format.mimetypeapplication/pdfen
dc.language.isoengen
dc.subjectLidaren
dc.subjectLight detection and rangingen
dc.subjectHydrologyen
dc.subjectTopographyen
dc.subjectElevationen
dc.subjectHydrological datasetsen
dc.subjectWater surface mappingen
dc.titlePoint cloud classification for water surface identification in Lidar datasetsen
dc.date.updated2011-07-07T19:50:05Zen
dc.identifier.slug2152/ETD-UT-2011-05-3151en
dc.contributor.committeeMemberMaidment, David R.en
dc.contributor.committeeMemberPassalacqua, Paolaen
dc.description.departmentCivil, Architectural, and Environmental Engineeringen
dc.type.genrethesisen
thesis.degree.departmentCivil, Architectural, and Environmental Engineeringen
thesis.degree.disciplineEnvironmental and Water Resources Engineeringen
thesis.degree.grantorUniversity of Texas at Austinen
thesis.degree.levelMastersen
thesis.degree.nameMaster of Science in Engineeringen


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record