Measuring discharge in channels with image processing techniques

Estimating surface velocity in hydraulic channels with image processing: experiments and validation with ADV

In this thesis you will turn videos recorded with a phone or a camera into data for measuring the surface flow velocity in channels. With image processing techniques (motion detection, particle tracking or computer vision algorithms) you will be able to analyze flow behavior in real time.

For the experiments you will have a hydraulic flume 4.5 m long and 30 cm wide for controlled tests. You will also measure velocity profiles in vertical and horizontal sections with an acoustic Doppler velocimeter (ADV), to correlate the image-based results with precise measurements.

The thesis is carried out with Universidad de Valparaíso, whose laboratories will host the experiments and initial validation, with a possible extension to drones in the field.

Possible applications include rivers, irrigation canals and drainage systems.

Objectives

• Develop and apply image processing algorithms to estimate surface velocities from video.
• Run experiments in the hydraulic flume and record data with cameras and the ADV.
• Analyze and correlate surface velocities with full flow profiles.
• Validate the method under varying conditions (for example, different flow rates and turbulence levels) and propose improvements for practical applications.