United States Department of Transportation
i
Integrated Analytics and Visualization for Multi-modality Transportation Data
-
2019-04-01
-
[PDF-2.66 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Resource Type:
-
Geographical Coverage:
-
Corporate Publisher:
-
Abstract:Urban informatics is an important field that is attracting increasing attention from academia, government and industry. One particular area of interest in this field is in modeling movement of people and goods around cities. In this report the authors leverage a new data set composed of tens of millions of images from New York City captured over a period of a year by cameras mounted on top of cars and produced by Brooklyn-based start-up Carmera. This data, by providing comprehensive coverage of not only the various streets of the city but also different time periods, has the potential to provide users with a visual perspective of the city that was not possible before. The authors first present a framework that leverages recent advances in computer vision to efficiently handle such a large collection of complex images. They then propose to construct a spatiotemporal map of relative pedestrian density. Due to the limitations of state-of-the-art computer vision methods, such automatic detection of pedestrians is inherently subject to errors. These errors are modeled as a probabilistic process and theoretical analysis is provided. Through numerical simulations, it is demonstrated that the methodology can supply a reasonable estimate of pedestrian densities and provide theoretical bounds for the resulting error. Lastly, an interactive visual analysis tool is presented for the exploration of this large collection of images. The approach computes a set of feature vectors for each image in this large street-level collection and makes use of a memory-efficient index to interactively answer queries about this data. A visual interface couples the image data with other urban data sets and allows users to interactively query, explore and analyze visuals in a city over both space and time. This provides them with the opportunity to not only virtually audit the built environment but also help answer numerous questions about the complex system of cities. Working in collaboration with urban planning researchers, the utility of the framework is illustrated through several use cases.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
