I-Corps: A machine learning and video-based sensor for measuring sewer flows

这个I-Corps项目更广泛的影响/商业潜力在于改善水回收作业以及对人类和环境健康的相关积极影响。美国的水回收设施每年向环境排放数十亿加仑未经处理的废水，原因是大风暴导致其收集系统溢出。这些系统的故障还可能导致地下室未经处理的水流入公民家中，这可能造成财产损失和严重疾病的风险。为了克服这些挑战，水回收设施依靠下水道流量数据来了解管网内发生的情况;然而，现有的传感器存在几个缺点，包括数据不准确，成本高，无法检测低流量。该技术旨在通过机器学习和基于视频的传感器来测量下水道流量和水质，从而克服这些挑战。这项技术有可能为水回收设施提供所有管道流量条件下准确可靠的数据，为昂贵的基础设施和运营决策提供信息。此外，通过帮助减少地下室备份，该技术具有更广泛的社会意义，因为低收入和少数民族社区受到洪水影响的影响不成比例。这个I-Corps项目是基于开发一种低成本的基于视频的传感器，使用机器学习和频谱分析来测量卫生下水道系统中的流量。具体而言，所提出的技术捕获下水道流量的视频，并使用机器学习算法来估计速度和水位。此外，它使用光谱分析来确定水的透明度和颜色，可用于识别污水系统中的污染物来源。这项技术的新颖之处在于，首次将基于视频的传感器应用于精确测量卫生下水道中的流量和水质。通过对污水管中水流的大小和质量提供更便宜、更可靠和更准确的数据，拟议的技术有可能推进水回收操作。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this I-Corps project is in the improvement of water reclamation operations and associated positive impacts on human and environmental health. Water reclamation facilities in the U.S. discharge billions of gallons of untreated wastewater into the environment each year due to large storms that cause overflows in their collection systems. Failures in these systems can also result in basement backups of untreated water into citizens' homes, which can cause property damage and risk of serious illnesses. To overcome these challenges, water reclamation facilities rely on sewer flow data to get a picture of what is happening within their pipe networks; however, existing sensors have several shortcomings including inaccurate data, high costs, and an inability to detect low flows. The proposed technology seeks to overcome these challenges through a machine learning and video-based sensor for measuring sewer flows and water quality. This technology has the potential to provide water reclamation facilities with accurate and reliable data across all pipe flow conditions to inform decision-making on costly infrastructure and operations. In addition, by helping to reduce basement backups, this technology has broader social implications as low-income and minority communities are disproportionately affected by flood impacts.This I-Corps project is based on the development of a low-cost, video-based sensor to measure flows in sanitary sewer systems using machine learning and spectral analysis. Specifically, the proposed technology captures video of sewer flow and processes using a machine learning algorithm to estimate velocity and water level. In addition, it uses spectral analysis to determine the clarity and color of water that can be used to identify sources of pollutants within the sanitary sewer system. This technology is novel in that for the first-time a video-based sensor is applied to accurately measure flow and water quality in sanitary sewers. The proposed technology has the potential to advance water reclamation operations through less expensive, more reliable, and more accurate data on the magnitude and quality of water flow in sanitary sewers.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.