Algorithm Speeds Lidar Assessment of Landslide Risk

Computer algorithms could turn lidar surveying systems into powerful tools for assessing landslides and keeping people out of their path.

Created by researchers at Oregon State University and George Mason University, the Contour Connection Method (CCM) is based on lidar data and requires minimal user input. The developers say it can analyze and classify landslide risk in an area of 50 or more square miles in about 30 minutes, a task that would otherwise take an expert several weeks to months to complete.

Current landslide-mapping techniques include field inventorying and photogrammetry to highlight regions of instability. Bare-earth lidar digital terrain models can reveal the landscape beneath vegetation and other obstructions, highlighting landslide features, including scarps, deposits and fans.

A computer algorithm detected this past landslide in the Stillaguamish Valley of Washington. Graphic courtesy of Oregon State University.

The new CCM algorithm looks for land features, such as suddenly steeper areas of soil, that might be evidence of a past landslide. It then searches the terrain for other features, such as a “toe” of soils at the base of the landslide. “Each landslide feature has a distinct set of metadata — specifically, density of connection vectors on each contour — that provides a unique signature for each landslide,” the researchers wrote in a study.

“With lidar we can see areas that are 50 to 80 percent covered by landslide deposits,” said Oregon professor Dr. Michael Olsen. “It may turn out that there are 10 to 100 times more landslides in some places than we knew of before.”

The technology was applied to the region surrounding the March landslide that killed 43 people near Oso, Wash. It analyzed more than 2200 acres in about nine minutes.

A faster way to assess landslide risk could enable better protection for homeowners.

“A lot of people don't think in geologic terms, so if they see a hill that's been there for a long time, they assume there's no risk,” said Oregon professor Dr. Ben Leshchinsky. “And many times they don't want to pay extra to have an expert assess landslide risks or do something that might interfere with their land development plans.”

Future versions of the technology could allow for real-time monitoring of soil movement, the researchers said.

The research was published in Computers and Geosciences (doi:10.1016/j.cageo.2014.10.007).

For more information, visit www.oregonstate.edu.