Authors: MELANIE LOPEZ*, San Diego State University
Topics: Remote Sensing, Land Use, Urban Geography
Keywords: OBIA, Machine Learning, Built Environment, Neighborhood Conditions
Session Type: Guided Poster
Start / End Time: 8:00 AM / 9:40 AM
Room: Roosevelt 3.5, Marriott, Exhibition Level
Presentation File: No File Uploaded
Neighborhood conditions are often determined by the availability and condition of its amenities. While the conditions have been determined through labor intensive field studies, we have determined that opportunities to assess conditions can be developed through advancements in remote sensing and machine learning. Aerial imagery from the National Agriculture Imagery Program can be processed with Google Earth Engine and the GSV API embedded in python scripts Further, built environment amenities such as amount of vegetation, buildings, swimming pools, and road networks can be detected using object based image analysis (OBIA). Although OBIA can successfully extract neighborhood features, this methodology can be time consuming given that spectral reflectance tolerances need to be adjusted on an image by image basis. Machine learning techniques incorporating Tensorflow can yield similar or even more accurate results with a more standardized approach across all images. This study compares the accuracy of the two methodologies.