During the 2020 COVID-19 pandemic, numerous research groups have worked to find effective ways to determine flow patterns and crowd densities on the streets of major cities like New York City to gain insight into the effects of stay-at-home and social distancing policies. But sending teams of researchers to the streets to observe and record the data exposes researchers to the risk of infection that these policies are designed to contain. New York University (NYU) Center for Smart Transportation Connected Cities Towards Accessible and Resilient Transportation (C2SMART), a U.S. Department of Transportation-funded Tier 1 transportation center, has developed a solution that not only eliminates the risk of infection for researchers , also provides easy access to existing public transit camera feed infrastructure, and also provides data on crowds and traffic density that is more comprehensive than previously compiled data and cannot be easily detected with traditional traffic sensors arrive.
To achieve this, researchers at C2SMART leveraged publicly available video feeds from more than 700 NYC Department of Transportation (DOT) locations, using a camera-based deep learning object detection method that allows researchers to Calculate pedestrian and traffic density. “The idea is to use these DOT cameras to feed and record, to better understand pedestrians’ social distancing behavior,” said Kaan Ozbay, director of C2SMART and a professor at NYU. Ozbay and his team wrote a “crawler”—essentially a A tool to automatically index video content – get low-quality images from video sources on the Internet. Each frame of the video is then processed using off-the-shelf deep learning image processing algorithms to learn what each frame contains: a bus, a car, a pedestrian, a bicycle, etc. The system also blurs identifiable images like all faces without compromising the effectiveness of the algorithm. The system can help policymakers understand a wide range of issues, from crisis management, such as social distancing behavior, to traffic congestion .
This article is reprinted from: https://www.solidot.org/story?sid=71823
This site is for inclusion only, and the copyright belongs to the original author.