Comparison of SfM computer vision point clouds of a landslide derived from multiple small UAV platforms and sensors to a TLS-based model

Abstract

Structure from motion (SfM) computer vision is a remote sensing method that is gaining popularity due to its simplicity and ability to accurately characterize site geometry in three dimensions (3D). While many researchers have demonstrated the potential for SfM to be used with unmanned aerial vehicles (UAVs) to model in 3D various geologic features, such as landslides, little is understood concerning how the selection of the UAV platform can affect the resolution and accuracy of the model. This study evaluates the resolution and accuracy of 3D point cloud models of a large landslide that occurred in 2013 near Page, Arizona, that were developed from various small UAV platform and camera configurations. Terrestrial laser scans were performed at the landslide and were used to establish a comparative baseline model. Results from the study indicate that point cloud resolution improved by more than 16% when using multi-rotor UAVs instead of fixed-wing UAVs. However, accuracy of the points in the point cloud model appear to be independent of the UAV platform, but depend principally on the selected camera and the image resolution. Additional practical guidance on flying various UAV platforms in challenging field conditions is provided for geologists and engineers.

Multitemporal monitoring of a coastal landslide through SfM‐derived point cloud comparison

“…Such deformation was also noted by Ruggles et al. () in the comparison of point clouds derived from different UAV/sensor combinations with a model derived from a terrestrial laser scanner. Those authors quantified the deformations through the C2C tool, which was also used in the current study.…”

“…It is worth mentioning that, although GoPro provides images with a large field of view and it can be easily carried by almost any UAV platform due to a weight of only 74 g, it is generally not recommended for accurate photogrammetric applications. The most important drawbacks consist of its fish‐eye lens producing curvature and distortion in SfM 3D reconstruction, and in a rolling shutter system for image acquisition (Thoeni et al., ; Ruggles et al., ). Conversely, the quality of images from DSLR cameras is always higher than a GoPro's, due to more stable optics, longer focal distances and higher‐quality sensors (charge‐coupled device (CCD) or complementary metal oxide semiconductor (CMOS)).…”

Accuracy of Photogrammetric UAV-Based Point Clouds under Conditions of Partially-Open Forest Canopy

“…RGB or NIR (near infrared) cameras are used in the majority of these studies. Many other applications of UAVs have also been described apart from forestry, e.g., agriculture [22,23], damage detection of buildings [24,25], landslide research [26], research of…”

Potential Benefits of Combining Anomaly Detection with View Planning for UAV Infrastructure Modeling

“…Together with the increasing ease of obtaining imagery, advances in computer vision and computer processing power have led to a widespread increase in aerial mapping and 3D-reconstruction [6][7][8]. In the field of infrastructure monitoring, the clear advantages of on-demand, high precision 3D modeling are driving companies and researchers to explore the possibilities of this technology (for an excellent overview, see [9]).…”