Rock-based Landslide Safety Monitoring Using Terrestrial Laser Scanning Technology
X-T. Guo and T. Huang
The one-dimensional (1-D) characteristics and insensitivity to small deformation are examples of the limitations that exist in the traditional methods for landslide safety monitoring using terrestrial laser scanning (TLS) technology. To counter such limitations, a novel approach was proposed based on multi-temporal scans, whose kernel relies on the use of a three-dimensional (3-D) matching algorithm to track and estimate the deformation of each analysis object (such as rocks) selected on the landslide, and mainly involves three steps: the acquisition of multi-period scans; the co-registration of the scanned data; and the parameter estimation based on local object matching. Due to this approach can adequately take advantage of the large redundancy in point clouds, it has the capability to provide precisely 3-D change information, including displacement and rotation, and has a high degree of flexibility that can be widely applied. By a simulation experiment, the performance of the approach was verified and some meaningful results were achieved: in case of scanning targets at a distance of 220 m, the estimated displacement errors are approximate 1 cm, the rotary errors are below 0.9°; the corresponding errors increase with the distance changed to 400 m. Additionally, following this approach a practical monitoring case of landslide was briefly analysed, and the analysis results further verified the findings obtained in the designed experiment.
Keywords: Terrestrial laser scanning (TLS), safety monitoring, landslide, rocks, algorithms, Gauss-Markoff model, parameters