Since 1992, coastal morphology and hydrodynamics of the nearshore zone have been studied from video observations, within the framework of the so-called ARGUS research program. Image data are collected every day-light hour, at seven beach locations worldwide. Timeaveraged images show bright, longshore bands of intensities, clearly indicating the locations where waves preferably break. In the nearshore zone waves generally break due to depth limitation. Because of this, locally observed light intensities can be assumed to be related to local bathymetry. This relationship has already been indicated qualitatively by Lippmann and Ho/man [1989]. In this thesis, the relationship between image intensities and bathymetry is quantified. A model called MONIMORPH ('MoNitoring MORPHOlogy') has been developed, which estimates the bottom elevation Zb from observed image intensities. This is performed by relating intensity values along a cross-shore transect to a wave parameter, and modelling this wave parameter inversely. For the time being, considerations are one-dimensional and concentrate on the actual region of wave breaking. From both statistical and physical considerations it was concluded that it might be useful to relate image intensities to the roller energy density E. divided by the squared phase speed c2 . In order to obtain a quantitative match between intensity profile and E/c2 curve, the raw intensity data are scaled by means of a three parameter model (Ibase , r, SF). The background intensity parameter Ibase and the trend removal parameter r are derived from raw image intensities, while the upscaling factor SF is related to the ratio Hsig/Hmax at the seaward boundary of the computational region. The MONIMORPH wave model comprises the inverted UNillEST-TC equations, UNlBEST-TC being a cross-shore morphodynamic model developed at DELFT HYDRAULICS. Based on boundary conditions for (Hnns' h, () and 11) and a cross-shore distribution of E/c2 , it computes the corresponding bottom elevation. Estimates obtained from single images are combined by means of a data assimilation technique. From a sensitivity analysis, a favourable mechanism, damping both initial disturbances in the boundary conditions and noise in the input intensity signal, was found to exist. It makes MONIMORPH suitable to deal with relevant initial deviations of the order of 5%. The inverse model has been calibrated against data obtained from the field campaign at Duck, October 1994, yielding a scaling relation for SF. Based on this relation MONIMORPH has been tested for 27 different situations. It was concluded that for situations within the range of calibrated wave conditions (the so-called calibration window), reliable estimates of bar bathymetry are produced: deviations at the top of the bar amount 10 to 20 cm, while the mean difference across the bar is 30 to 40 cm. The bar crest is systematically predicted too far shoreward, though the differences of 10 to 20 m are small considering the mild slope of the bar. Situations not matching the wave window have to be excluded from analysis, for the time being.