Automatic extraction of yardangs using Landsat 8 and UAV images: A case study in the Qaidam Basin, China

Yardangs are ridge features of various scales indicating the direction of prevailing wind when they evolved. In this paper, we apply an integrated approach for the automatic extraction of yardangs in typical areas of the Qaidam Basin. Canny's edge detection algorithm is used for yardangs in 15-m Landsat 8 Operational Land Imager (OLI) data to acquire the spatial distribution patterns of yardangs, and ellipse-fitting algorithm is used to get the morphological parameters of yardangs in 1.7-m Digital Elevation Model (DEM) generated by unmanned aerial vehicle (UAV) images. Both the spatial distribution patterns and morphological parameters can improve knowledge of yardangs. The experimental results illustrate the accuracy and feasibility of the proposed extraction approach for the spatial distribution patterns of yardangs with an average relative error no more than 10%. The statistical analysis shows that most of yardangs developed near anticlines and few in synclines; most yardangs are in Pliocene strata, and a few in lower and middle Pleistocene strata. For the automatically extracted morphological parameters, the relative error of each parameter is less than 6%, which is meaningful for the type identification of yardangs.