Abstract:
Objectives Terrestrial laser scanning (TLS), ground-based interferometric synthetic aperture radar (GB-InSAR) and unmanned aerial vehicle(UAV) photography have been applied to emergency deformation monitoring of rockslide. The objectives are concerns about new problems of 3D visualization of deformation data.
Methods Firstly, we introduced iterative closest point method (ICP) to complete the registration between the TLS point cloud and UAV terrain model. Secondly, geometric mapping method was applied to achieve data fusion between GB-InSAR deformation map and TLS point cloud.The mapping deviation is difficult to be corrected by conventional methods, because the lack of artificial control points on the rockslide. Visual interpretation and GB-InSAR images simulation method were combined to extract the control points.The spatial coordinate transformation parameters were estimated with the least square method. The transformation model was established to correct the mismatch.
Results The proposed method was verified by the simulation data and the actual monitoring data of a landslide. The result shows that the accuracy reaches the pixel level which meet the needs of emergency monitoring.
Conclusions The proposed method is limited by the working conditions of emergency and the subjectivity of image interpretation. If the accuracy of control point selection was not narrowed, the final matching fusion position accuracy may reach pixel-level accuracy.
下载:
