A Low-Cost Modular Image-Based Approach to Characterize Large-Field Wave Shapes in Glass Wave Flume

Abstract

Accurately determining water surface elevation and wave shapes in the hydraulic laboratory is critical for experimental research and physical understanding of ocean waves. Existing technologies such as wave gauges cannot capture the continuous wave profile across both space and time. This poses an issue, as nonlinear wave characteristics vary as a function of position and cannot be fully described using such point measurements. Furthermore, wave gauges are intrusive to the flow field. Alternative single-camera methods can’t capture wave characteristics in a large field-of-view properly without sacrificing resolution. In this paper, the authors propose an easy-to-use, low-cost method for measuring wave height and shape along the length of the flume over time. The method utilizes stitching of multiple web-cameras and the application of a Canny-based edge detection algorithm with experimentally determined thresholds and additional filters for maximum robustness and efficacy. Additionally, distortion correction is implemented in a computationally efficient manner. Video is acquired by three Logitech C920 PRO HD cameras recording at a resolution of 1280 × 720 at 24fps. The wave generator can generate waves with frequency between 0.1Hz and 1Hz. The experimental results show that wave height measurements can be obtained with a maximum resolution of 0.83mm with a relative error of ±1.5% when compared with a reference wave gauge measurement. This work demonstrates the ability to arbitrarily extend the horizontal field-of-view while providing more accurate measurement results.