Robust and reliable information is needed regarding stormwater infrastructure performance to effectively manage pollutants from stormwater runoff. In particular, measuring outlet flow magnitudes is important as outlets provide the critical function of regulating flow for structural stormwater control measures (SCMs), thus outlet monitoring aids in evaluating performance and identifying possible design problems and causes for failure. The current state of practice is to use sensors or other instrumentation for monitoring flow; however, these systems can be expensive, require multiple components, and are placed in harsh environments, which decreases their expected lifetime. Modeling approaches may be used instead to predict flow; however, model accuracy relies on rainfall estimates and information regarding SCM conditions, which may not always be available or accurate. Therefore, there is a need to develop methods for low-cost, quick-deployment, extended-term outlet monitoring such that timely action can be taken when problems first occur, to evaluate performance during significant precipitation events, and to have a better understanding of system performance over time. To address this need, this project aims to implement and validate a long-term, low-cost, and accurate computer vision-based technology to monitor pipe outlet flow in the field.
To achieve this goal, the following objectives will be pursued: 1) create a typology of cases in the field of pipe outlet flow (flush and perched), from which 2) a computer vision method for measuring pipe outlet flow will be developed and validated; 3) perform field validation of a computer vision method for measuring outlet flow pipe structure; 4) evaluate the effects of outlet pipe type and pipe material on flow measurement algorithm performance. We will build on prior work that developed a computer vision method for estimating flow from a perched pipe in a controlled environment by validating this technology in the field. Simultaneously, we will
develop a computer vision method to estimate pipe flow from a flush pipe structure using background subtraction, motion detection, and particle tracking, followed by in-field validation. Finally, we will conduct a long-term field study to evaluate the effects of pipe type (perched, flush) and material (smooth, corrugated) on flow estimation algorithm performance. Flow rate is an important yet demanding parameter to measure within the fields of hydrology and hydraulics. This project will develop a low-cost, non-contact method that uses a single device to estimate flow using newly developed and validated image processing methods, providing clear advantages over expensive and intrusive installations with the added flexibility of a single
component system. This project will also generate a set of recommendations and best practices, presented in the form of a written document, for installing camera systems for monitoring stormwater outlet infrastructure, including camera hardware specifications, camera installation considerations, and data management practices, for successful technology transfer to NCDOT.
