Satellite derived photogrammetric bathymetry

Satellite Derived Bathymetry (SDB) is being adopted as a cheaper and more spatially extensive method for bathymetric mapping than traditional acoustic surveys, with research being conducted by the Canadian Hydrographic Service under a Government Related Initiatives Program (GRIP) of the Canadian Space Agency. Established SDB methods involve either an empirical approach, where a regression between known depths and various colour indexes is developed; or a physics-based Radiative Transfer Model (RTM) approach, where light interactions through the water column are simulated. Both methods have achieved vertical accuracies of around 1 m. However, the empirical approach is limited to areas with existing in-situ depth data, and has limited applicability in heterogeneous benthic environments, while the physics-based approach requires precise atmospheric correction. This paper proposes a through-water photogrammetric approach which avoids these limitations, in heterogeneous seafloor environments, by using feature extraction and image geometry rather than spectral radiance to estimate bathymetry. The method is demonstrated in Coral Harbour, Nunavut, Canada using a WorldView-2 stereo pair. A standard photogrammetric extraction was performed on the stereo pair, including a blunder removal and noise reduction. Apparent depths were then calculated by referencing under-water points to the extracted elevation of the water-line. Actual in-image depths were calculated from apparent depths by applying a correction factor to account for the effects of refraction at the air-water boundary. A tidal reduction brought depths to local chart datum, allowing for validation with Canadian Hydrographic Service survey data showing a mean error of 0.031 m and an RMSE of 1.178 m. The method has a similar accuracy to the two established SDB methods, allowing for its use for bathymetric mapping in circumstances where the established methods are not applicable due to their inherent limitations.