Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Land Surveying and Geo-Informatics | en_US |
| dc.contributor.advisor | Ding, Xiaoli (LSGI) | en_US |
| dc.creator | Shahzad, Naeem | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/12198 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Modeling landslide hazards through susceptibility analysis and slope deformation measurements | en_US |
| dcterms.abstract | The focus of this research is three-fold; (1) landslide susceptibility mapping with machine learning models and their performance evaluation in northern Pakistan, (2) wide area processing (WAP) for mapping active deformation areas (ADA) in northern Pakistan with integrated persistent scatterer interferometry (PSI) approach, and (3) investigating landslides with an improved multi-temporal InSAR (MT-InSAR) approach through ALOS/PALSAR-2 data. At a very first stage, performances of five machine learning techniques, including random forest (RF), support vector machine (SVM), maximum entropy (maxENT), gradient boosting machine (GBM), and logistic regression (LR), were investigated for landslide susceptibility mapping (LSM) in the rugged terrain at regional scale in northern Pakistan with available landslide inventory information of 200 samples. For this purpose, 12 landslide influencing factors (LIFs) were selected and scanned through the two different tests based on information gain ratio (IGR) and multi-collinearity before modeling the susceptibility and their performance evaluation measures, including calculation of area under the curve (AUC), parameters of the confusion matrix and overall performance (POA). The results showed that the SVM was the most promising model (AUC =0.969, POA = 2669) for the LSM followed by RF (AUC =0.967, POA = 2656), GBM (AUC =0.967, POA = 2623), maxENT (AUC = 0.872, POA = 1761), and LR (AUC = 0.836, POA = 1299). | en_US |
| dcterms.abstract | PSI technique faces challenges when applied to large mountainous regions in terms of low-density PS, its distribution, and other challenges related to computational burden and processing time when performing wide-area processing (WAP). Therefore, at a second stage, an integrated approach was introduced to process large extent SAR images and applied to multi-orbit Sentinel-1A data for mapping active deformation areas in northern Pakistan. The integration has been done at three different stages including utilizing three different PS approaches (i.e., spectral diversity, temporal variability, and phase stability), conceptualizing an overlap of 60 percent between the divided blocks of a wider SAR image, and bridging the final PSI based deformation product with GIS-based strategies for mapping active deformation areas (ADA). Considering the benefits of working with vector data sets of interferometric point target analysis (IPTA), each block is processed separately for deformation mapping and merged using a block adjustment algorithm. The final LOS deformation results are processed through a couple of post-processing steps, including translation into vSLOPE, the use of Density-based spatial clustering of applications with noise (DBSCAN) to detect and classify the ADAs into active deformation slopes (ADS), erosion, and subsidence. The integration has been found to produce smoothed and continuous deformation results with LOS velocity ranges from -60 mm/year to +60 mm/year also helped to identify 2500 ADS of varying sizes upto 1 sq. km and PS density of 58 PS/sq.km and 65 PS/sq.km in ascending and descending orbit data respectively. In the presence of serval natural and anthropogenic factors, the causes of deformation in the area cannot be directly linked. However, the distribution analysis showed that ADS exhibited close relationships with distances to faults, drainages, roads, and precipitation as a speeding tool. Taking advantage of significant overlap, the results generated from this tested approach demonstrated its effectiveness interms of the smoothness of the final deformation map and time consumption (3.3 percent lesser time) with limited system resources when compared with the results generated from the standard PSI approach. The results also highlighted the effectiveness of using Sentinel-1 data in detecting a large number of unstable slopes at this scale in rugged mountains when compared to the standard approach. | en_US |
| dcterms.abstract | To further utilize the high density of point targets, an improved MT-InSAR approach with joint processing of PS and distributed scatterer (DS) is tested at the third stage. The Fast Statistically Homogeneous Pixel Selection (faSHP) method has been used to detect DS and processed jointly with PS in the interferometric point target analysis (IPTA) framework. The improved method is employed to investigate the unstable slopes in the Muzaffarabad region of northern Pakistan using 16- ALOS/PALSAR-2 images acquired between October 2014 and May 2020. With a point density of 11,707/ km2, the results obtained by implementing this method revealed the existence of dozens of unstable slopes with an average deformation rate varying between -40 mm/year and 20 mm/year along the line of sight (LOS) direction. The study demonstrates the effectiveness of the proposed method with approximately four (4) times more points targets than conventional PSI helped in achieving good results and properly analyzing the changing deformation pattern and its extent within the deforming slopes in this region. Identifying the re-activation of several other slopes in the region, this research also reports the discovery of a giant destabilized slope near the Patika village along the Neelam River, which extends about an area of 2.30 sq. km. Overall, the research demonstrated the effectiveness of the standard approaches with higher point densities and lesser time consumption, providing smoothed and accurate deformation results. The research approaches used in this study and its findings related to its proneness and prior detection of unstable/deforming areas shall help the decision and policy-making bodies and provide an essential guideline for effective mitigation measures. | en_US |
| dcterms.extent | xiv, 151 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2022 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.LCSH | Slopes (Soil mechanics) | en_US |
| dcterms.LCSH | Landslides -- Prevention | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | open access | en_US |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12198