Presentation Title

LANDSLIDE DETECTION ALONG TRANSPORTATION NETWORKS USING SPATIAL DATA

Start Date

November 2016

End Date

November 2016

Location

HUB 265

Type of Presentation

Oral Talk

Abstract

A landslide is defined as a mass of rocks and soils sliding down slopes of a mountain or a hillside. It causes damages on infrastructure and impacts human lives. In addition, landslides prolong commute time and delay services along transportation networks. For these reasons, it is imperative to detect early-warning signs of landslide hazards to mitigate maintenance cost and diminish the influence on human life. In this study, we present a methodology based on k-means clustering that utilizes a LiDAR-derived DEM to extract spatial features of landslides. The method tests several feature extractors to determine if the delineation between slide and non-slide terrain is possible. To evaluate the performance of the method, a test data set from the Carlyon Beach Peninsula in the state of Washington was evaluated. Initial results indicate that 80% of the landslide area was correctly identified when compared to the independently compiled landslide inventory map. These results suggest that the proposed methodology and DEMs can be used to detect and map landslides.

This document is currently not available here.

Share

COinS
 
Nov 12th, 11:15 AM Nov 12th, 11:30 AM

LANDSLIDE DETECTION ALONG TRANSPORTATION NETWORKS USING SPATIAL DATA

HUB 265

A landslide is defined as a mass of rocks and soils sliding down slopes of a mountain or a hillside. It causes damages on infrastructure and impacts human lives. In addition, landslides prolong commute time and delay services along transportation networks. For these reasons, it is imperative to detect early-warning signs of landslide hazards to mitigate maintenance cost and diminish the influence on human life. In this study, we present a methodology based on k-means clustering that utilizes a LiDAR-derived DEM to extract spatial features of landslides. The method tests several feature extractors to determine if the delineation between slide and non-slide terrain is possible. To evaluate the performance of the method, a test data set from the Carlyon Beach Peninsula in the state of Washington was evaluated. Initial results indicate that 80% of the landslide area was correctly identified when compared to the independently compiled landslide inventory map. These results suggest that the proposed methodology and DEMs can be used to detect and map landslides.