The exploration of mechanisms by which cinema contributes to our understanding of cities has interested both cinematographers and urban designers. Urban Visions explores the advancements in the field of computer vision to not only seek alternative cinematic documentation techniques, but also to explore quantitative image classification methods which could benefit environmental studies. This research catalogs the application of computer vision techniques and image classification algorithms from the largest to smallest scales applicable to the fields of architecture, urban design and environmental sciences. 

Starting from the large scale, for the past 40 years the field of Remote Sensing has revolutionized our understanding of environmental data acquisition. In the past two decades advancements in computer science have allowed us to develop sophisticated image classification algorithms for not only environmental hazard detection, but also object detections. This research focuses on working with the LandSat8 database to create a series of studies of the Los Angeles County area in order to detect various urban patterns such as the green vs. built area. Zooming into the building scale, a series of lidar scanning coupled with thermal imaging studies have allowed for material detection analysis. Zooming in further on the building components, a number of object detection techniques are tested in order to begin classifying building attributes such as windows, entrances, and roof types. The aim is to train a database of urban and building components that will allow one to input images from the macro to micro scale.