Visualization of Urban Prediction Using Augmented Maps

Abstract

In this paper, we explore a visualization method using augmented maps for urban prediction. Our implementation allows users to determine the location for prediction in a paper map. As an application example, we examine an area before and after new train station is built. We use the difference between two maps for simulating the changes or predicting the impact if a new train station is built on a location in a paper map. In off-line phase, we gather knowledge data from several reference locations by comparing two aerial maps (before and after the train station is built). We then analyze the difference of green spaces between those two maps using color extraction. We observe that the green space around the new train station mostly decreases due to the area development. This information is then stored for prediction reference. In on-line phase, we use a monocular setup that consists of one camera and a monitor display. A paper map is captured using a web camera and tracked using its geometrical features. These features can be provided using the available data from Geographical Information Systems (GIS) or automatically extracted from the texture. The map is then matched with the reference map in database. When the map is matched, we can overlay the simulation on how the green space will change due to the existence of new train stations on a new location inputted by the user.