Using Pre-Existing Cameras to Monitor Coastal Geophysical Processes With Open-Source Tools

Abstract:

Shore-based video cameras are widely-used to study coastal geophysical processes. However, research-grade camera systems within the Argus and other networks are deployed at a limited number of sites worldwide. In contrast, an extensive network of non-research web-cameras are operating on coastlines globally and live streaming their video to the web. These "surf-cameras" (surfcams) represent a mostly untapped source of video-image data for coastal research due to unknown intrinsic and extrinsic camera parameters and logistical concerns such as acquisition of ground control points (GCPs). The newly-developed and open-source Surf-camera Remote Calibration Tool (SurfRCaT) software was designed to exploit this existing infrastructure by automatically completing camera calibrations and image rectifications using remotely-extracted GCPs derived from airborne lidar observations within a novel photogrammetric workflow. In this study, we use SurfRCaT to examine coastal geophysical processes from two pre-existing surfcams on the Atlantic coast of Florida. Results show that SurfRCaT-derived image products are of accuracy consistent with those created from field measurements and existing Coastal Imaging Research Network (CIRN) image processing routines, with root mean square reprojection errors below 6 m at both cameras. SurfRCaT-derived products are being used to quantify changes to intertidal bathymetry and sandbar shape at these sites, with all analysis routines (e.g. automatic shoreline extraction) written in Python as part of an initiative to translate existing CIRN Matlab routines. This study exemplifies SurfRCaT's ability to transform existing coastal cameras into quantitative monitoring tools, illustrating its potential to expand the number of coastal sites that can be studied via video-imagery.