River Flow Monitoring: The Influence of Graphical Enhancement Techniques on Image Velocimetry Performances

Abstract

Image velocimetry techniques for river flow monitoring have progressively spread due to the advantages of such methods with respect to traditional approaches. Image-based techniques can nonintrusively provide river discharge, with accuracy depending on several factors (e.g., environmental and hydraulic conditions, processing software, etc.). A key stage of image-based techniques workflow is frames pre-processing. In particular, graphical enhancement is often needed to maximize contrast between tracer and background, improving software capability to detect and track the tracer motion. This work aims to investigate the influence of graphical enhancement methods on the results of image-based analyses, comparing traditional and less common algorithms. Analyses were conducted on videos acquired during a measurement campaign on two rivers in Sicily (Italy), where simultaneous ADCP reference measurements were also collected. Different graphic filters were applied to frames and the enhanced sequences were then processed using two optical software programs (i.e., PIVlab and SSIMS-Flow). Performances are evaluated in terms of errors in surface velocity and discharge assessment, using ADCP measurements as benchmark. Results confirm the overall high potential of image-based techniques and provide insights on the importance of using the appropriate enhancement filters based on the selected processing software and environmental conditions at recording time.