Author : Xin Hong
Publisher :
ISBN 13 :
Total Pages : 0 pages
Book Rating : 4.:/5 (139 download)
Book Synopsis A Convolutional Neural Network for Detecting and Mapping Built Environment at Neighborhood Scale by : Xin Hong
Download or read book A Convolutional Neural Network for Detecting and Mapping Built Environment at Neighborhood Scale written by Xin Hong and published by . This book was released on 2021 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: The increasing interest in the connection between built environment and health has encouraged the development of new tools for describing health-related built environment. With the development of deep learning, scholars have been exploring the applications of convolutional neural networks (CNNs) to the field of remote sensing. Nevertheless, applying deep learning in remote sensing is still a young field. Instead of treating deep learning as a "black-box" technology, this study embraced deep learning as the key to solving large-scale and high-resolution remote sensing scenes. This study applied U-net, an encoder-decoder CNN architecture, for detecting greenness at street level. A new operational definition of the concept of neighborhoods: sidewalk-homogenous neighborhoods, which corresponds to different economic levels and habits of using sidewalks, was also proposed as a novel and practical delineation of neighborhood boundaries. As a pilot study, this study tested that deep learning is a sufficient method for detecting built environment on high volume unmanned aerial vehicle (UAV) images. The sidewalk-homogenous neighborhoods is a reasonable spatial scale that can help to reveal the disparities in sidewalk environments between neighborhoods.