Swell-dominated wave parameter retrieval from oblique coastal video under variable camera geometry

Nearshore wave monitoring remains constrained by the cost and logistical complexity of in-situ instrumentation, creating persistent observational gaps across coastal environments. This study introduces a hybrid single-camera framework for estimating significant wave height (Hs) and peak wave period (Tp) from oblique optical imagery of shoaling waves, without ground control points, calibration targets, or auxiliary sensors. The method integrates vanishing-point geometry, solar-elevation and azimuth shadow correction, and frame-wise spectral decomposition of pixel brightness signals. Two Hs methods are implemented: a frame-by-frame shadow-based estimate and a wave-cycle-resolved estimate from surface elevation reconstruction. Direct validation is conducted at Praia de Mira on the Portuguese west coast using UAV and fixed-camera deployments with co-located pressure sensor measurements. Tp estimates achieve MAPE below 10% and the frame-by-frame Hs method outperformed the surface elevation reconstruction by 5.9% in relative error and was adopted as the primary Hs estimator for the fixed camera deployments. The framework is subsequently demonstrated under sensor-absent conditions at the Port of Sines (approximately 1700 m standoff distance), where over 80% of Tp estimates fall within the operational benchmark and robust Hs retrieval is achieved predominantly for wave heights exceeding 1.5 m under the present deployment geometry. Retrieval performance is governed by optical geometry, illumination conditions, and wave climate within a swell-dominated shoaling operating domain. These findings establish a physically constrained optical retrieval framework supporting low-cost single-camera wave monitoring across data-scarce coastal settings.