Mapping of elements at risk for landslides in the tropics using airborne laser scanning

Khamarrul Azahari Razak, C.J. van Westen, Menno Straatsma, ... [et al.]

Research output: Contribution to conferencePaperAcademicpeer-review


Mapping elements at risk for landslides in the tropics pose as a challenging task. Aerial-photograph, satellite imagery, and synthetic perture radar images are less effective to accurately provide physical presence of objects in a relatively short time. In this paper, we
utilized an airborne laser scanning (ALS) for extracting elements at risk for landslides, which we emphasize d on the buildings and roads extraction in a populated tropical region (Cameron High lands, Malaysia). W
e presented the building filter derived from the hierarchical robust interpolation method for building extraction. Meanwhile,the road extraction as performed based on the ALS-derived topographic openness, analyzed inastereoscopic model. Building and road attributes in relation to landslides were ubsequently generated such as perimeter and area of building footprint;number nd height of the buildings;road location;length;road gradient, and road-cuts. We quantitatively evaluated the building detection method and measured the vertical accuracy of ALS-Derived road. The evaluation showed he building detection rate of 88.6%, the correctness of 90% and the overall uality f 80.7%. The vertical accuracy of the ALS-derived road wasabout0.68 mand spatially improved compared to the existing road map. This study illustrates the effectiveness of ALS data for mapping elements at risksin the tropics, which are ssential for landslide vulnerability and risk assessment.
Original languageEnglish
Number of pages16
Publication statusPublished - 2011
EventFIG Working Week 2011: Bridging the Gap between Cultures - Marrakech, Morocco
Duration: 18 May 201122 May 2011


ConferenceFIG Working Week 2011

Fingerprint Dive into the research topics of 'Mapping of elements at risk for landslides in the tropics using airborne laser scanning'. Together they form a unique fingerprint.

Cite this