Landslide assessment by using multi-temporal UAV datasets: a case study in northern Pakistan

Mountainous areas with activity seismicity are often prone to frequent and widespread landslides. Assessment and monitoring of landslides using coarse resolution spaceborne sensors limit the identification of minor landslide features and evaluating ground displacement. UAVs are emerging as an effective tool for assessing and monitoring landslides at a large scale with a very high spatial resolution. This study evaluates the landslides displacement using UAVs derived aerial photos, digital surface model (DSMs), and field data, in the Nagar valley, northern Pakistan. The orthomosaic and DSMs were computed using GCPs for orthorectification in the Pix4D Mapper and the dense point cloud. Analysis of the derived DSMs shows that the volumetric changes on the landslide are mainly concentrated on the scarp and lateral portions of the slide. Image correlation in the COSI-Corr shows the ground displacement of 1 to 12 m. The developed methodology can be effectively applied for the assessment and monitoring of high-risk landslides in other regions.