Filtering of Single Photon LiDAR Data
| dc.contributor.advisor | Glennie, Craig L. | |
| dc.contributor.committeeMember | Shrestha, Ramesh L. | |
| dc.contributor.committeeMember | Starek, Michael J. | |
| dc.contributor.committeeMember | Lee, Hyongki | |
| dc.contributor.committeeMember | Hartzell, Preston J. | |
| dc.creator | Wang, Xiao | |
| dc.creator.orcid | 0000-0001-6242-3816 | |
| dc.date.accessioned | 2018-03-12T17:56:12Z | |
| dc.date.available | 2018-03-12T17:56:12Z | |
| dc.date.created | December 2017 | |
| dc.date.issued | 2017-12 | |
| dc.date.submitted | December 2017 | |
| dc.date.updated | 2018-03-12T17:56:12Z | |
| dc.description.abstract | Light detection and ranging (LiDAR) is a powerful technique that provides accurate three-dimensional measurement of targets. New generation single photon LiDAR (SPL) systems are able to collect data at faster rates than conventional linear-mode LiDAR and at a lower cost. However, the new SPL systems are also very sensitive to false returns, which cause a high noise rate in the collected data. As a result, SPL systems present new data processing challenges. The filtering of SPL data has different requirements than imaging or conventional linear-mode LiDAR data. There is no intensity information for each return, and the data structure of the 3-D point cloud is irregularly spaced and different than 2-D images, therefore most imaging processing methods cannot be applied directly to SPL data. It also has a much higher noise rate than conventional LiDAR data, which makes conventional LiDAR filtering methods inapplicable. There has been some initial research on techniques for filtering SPL data; however, the problem has not been fully studied. In this dissertation, new SPL filtering algorithms which accommodate the higher noise and false return rates, along with afterpulsing will be developed and analyzed. First, two novel filtering methods are proposed to remove solar/dark noise for 2-D and 3-D SPL data respectively. The 2-D method considers inhomogeneous noise distribution and local point distribution to provide a more reliable filtering result. The 3-D method utilizes principle component analysis to remove near-signal noise more effectively. Results from these two methods were compared with current methods on different types of terrain. The proposed methods removed more noise points and had lower RMSE compared to the reference data. An improved version of the proposed 3-D filter that better retains linear features such as powerlines is then proposed and validated. Finally, a method based on robust regression is proposed to remove afterpulses. We found that on average, 90% of the afterpulse points were removed over rooftop areas and the mean elevation difference with respect to a reference surface was reduced from 1.95 m to 0.23 m. | |
| dc.description.department | Civil and Environmental Engineering, Department of | |
| dc.format.digitalOrigin | born digital | |
| dc.format.mimetype | application/pdf | |
| dc.identifier.uri | http://hdl.handle.net/10657/2854 | |
| dc.language.iso | eng | |
| dc.rights | The author of this work is the copyright owner. UH Libraries and the Texas Digital Library have their permission to store and provide access to this work. Further transmission, reproduction, or presentation of this work is prohibited except with permission of the author(s). | |
| dc.subject | Single photon LiDAR | |
| dc.subject | Filtering | |
| dc.title | Filtering of Single Photon LiDAR Data | |
| dc.type.dcmi | Text | |
| dc.type.genre | Thesis | |
| local.embargo.lift | 2018-12-01 | |
| local.embargo.terms | 2018-12-01 | |
| thesis.degree.college | Cullen College of Engineering | |
| thesis.degree.department | Civil and Environmental Engineering, Department of | |
| thesis.degree.discipline | Geosensing Systems | |
| thesis.degree.grantor | University of Houston | |
| thesis.degree.level | Doctoral | |
| thesis.degree.name | Doctor of Philosophy |
Files
Original bundle
1 - 1 of 1