An approach for monitoring landslide deformation using least-square matching algorithm

In application of the current Terrestrial Laser Scanning (TLS) technique suffers several limitations, such as its 1-D nature in monitoring the deformation and insensitivity to infinitesimal deformation. Here a new approach based on the data obtained from repeatedly scanning is proposed for monitoring the slope deformation. The new approach includes three main steps: sampling the raw TLS data, registering the obtained TLS data, and computing the deformation based on the subareas matching method. Central to the proposed procedure is a 3D matching algorithm (a generalized concept of the least-squares 2D image matching). Taking advantage of the high redundancy of TLS data, the procedure can be used to monitor the full 3-D deformations with high resolution, yielding the transitional and rotational displacements of the monitored objects, while possessing high flexibility in application. The performance and advantage of the approach are illustrated through a designed simulation experiment. In addition, a case study is performed based on the proposed approach, further confirming the result of the simulation experiment. This procedure is applicable to the deformation monitoring, especially for the remote areas, helping promote the transition of the current deformation monitoring method from pointwise to morphological monitoring.