Please use this identifier to cite or link to this item:
Title: Local-measure-based landslide morphological analysis using airborne LiDAR data
Authors: Deng, Susu
Keywords: Landslides -- China -- Hong Kong.
Landslides -- Risk assessment.
Remote sensing.
Hong Kong Polytechnic University -- Dissertations
Issue Date: 2014
Publisher: The Hong Kong Polytechnic University
Abstract: Landslides leave discernable signatures related to form, shape and appearance of land surface, i.e. morphological features, which are important for analysis of landslide mechanism, activity state and landslide detection. For objective analysis of landslide morphology, an approach capable of providing quantitative expression of morphological features is required. In addition, the development of new technique, namely airborne Light Detection And Ranging (LiDAR), allows morphological analysis in great details. Therefore, the approach should be able to express landslide morphological features related to multiple scales. Despite a number of methods based on mathematical tools have been developed to highlight particular information associated with landslide morphology, few efforts were devoted to quantification of landslide morphological features based on their descriptions. In this research, an approach based on local measures of spatial association was developed to quantify landslide morphological features represented by dominant morphology or topographic variability in a particular pattern. The use of local measures enables quantification of distinctness of landslide morphological features in a statistical way so that distinct morphological features can be identified. For characterization of spatial patterns of topographic variability, a method constructing local measure plots was proposed. A local measure plot indicates scales and magnitudes of topographic variations along a specified direction. Multi-scale patterns of topographic variability can be revealed based on the plots. Due to its capability of identifying landslide morphological features, the local-measure-based approach can be applied to landslide detection. In related researches of automated landslide detection, morphological features have not been thoroughly exploited, especially for detection of debris slides and flows. Thus a semi-automated landslide detection approach based on morphological features was proposed to identify locations of small size, shallow debris slides and flows. Landslide component candidates were extracted by identifying morphological features using the local-measure-based approach. Geometric and contextual analyses were subsequently conducted to discriminate landslide components from other terrain objects. The approach was applied to a test site containing both new and old landslides covered by dense vegetation. Owing to the vegetation penetration ability, airborne LiDAR was utilized. Almost all (93.6%) the new and a part (23.8%) of old landslides with distinct morphological features were detected. In this research, airborne LiDAR data was employed to generate high-resolution Digital Terrain Models (DTMs) utilized in landslide morphological analysis and landslide detection. Since land surface analysis is affected by the quality of DTM , the possibility of improving the filtering results of LiDAR point cloud using an integration scheme was explored. Through visual evaluation of the integration result, this scheme was proved to be able to remove a part of filtering errors and increase the number of ground points.
Description: xii, 158 pages : color illustrations ; 30 cm
PolyU Library Call No.: [THS] LG51 .H577P LSGI 2014 Deng
Rights: All rights reserved.
Appears in Collections:Thesis

Files in This Item:
File Description SizeFormat 
b2762836x_link.htmFor PolyU Users203 BHTMLView/Open
b2762836x_ir.pdfFor All Users (Non-printable)7.91 MBAdobe PDFView/Open
Show full item record
PIRA download icon_1.1View/Download Contents

Page view(s)

Last Week
Last month
Citations as of Oct 21, 2018


Citations as of Oct 21, 2018

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.