Measurements of short water waves using stereo matched image sequences

Abstract

Image analysis techniques are used for retrieving water surface elevation fields spatially and temporally from CCD-images and CCD-image-sequences. The technique proposed herein utilizes binocular stereogrammetry to recover topographic information from a sequence of synchronous, overlapping video images. The method used differs from the traditional stereo-photogrammetric analysis of a single stereo-pair because the use of video allows for a continuous sequence of stereo-images to be digitally sampled and analyzed. For data acquisition two synchronized progressive-scan cameras were used. A partially supervised 3-D stereo system (called WASS, Wave Acquisition Stereo System) is shown here. It is used to reconstruct the 3-D shape of water surface waves, acquired at frame rate, with small computational time needed. The stereo method is presented, including the derivation of a relationship relating the geometry of the stereo rig and the expected errors. Finally, the 3-D calculated scattered points give the complete spatio-temporal distribution of the water surface elevations. The measurable length-scales depend on the pixel resolution, the triangulation accuracy, and the acquisition frame rate. Limitations in the stereo measurements are also discussed. Two experiments to test and to demonstrate the system took place: one on the Venice lagoon, north of the city of Venice in September 2004 and the second on the coast of California at San Diego in December 2005. For the second experiment, qualitative and quantitative intercomparisons of the stereo-matching and in-situ sensor measurements are presented. All the measurements of water surface waves indicate that the proposed approach is both accurate and applicable for measuring water surface waves. Moreover, shape estimates are accurately and extremely dense both in space and time, and the remote location of the instruments alleviates some difficulties associated with in situ instrumentation.