Abstract
Abstract. In this paper we propose an approach to multi-class semantic segmentation of urban areas in high-resolution aerial images with classspecific object priors for buildings and roads. What makes model design challenging are highly heterogeneous object appearances and shapes that call for priors beyond standard smoothness or co-occurrence assumptions. The data term of our energy function consists of a pixel-wise classifier that learns local co-occurrence patterns in urban environments. To specifically model the structure of roads and buildings, we add high-level shape representations for both classes by sampling large sets of putative object candidates. Buildings are represented by sets of compact polygons, while roads are modeled as a collection of long, narrow segments. To obtain the final pixel-wise labeling, we use a CRF with higher-order potentials that balances the data term with the object candidates. We achieve overall labeling accuracies of > 80%.
Highlights
The automatic interpretation of aerial images has been a classic problem of remote sensing and machine vision
Interpreted images, i.e. thematic raster maps, of urban areas are important for many applications, for example mapping and navigation, urban planning and environmental monitoring, to name just a few
In spite of great progress, the task is far from solved. This is especially true for urban areas, and at high spatial resolutions
Summary
The automatic interpretation of aerial (and satellite) images has been a classic problem of remote sensing and machine vision. Interpreted images, i.e. thematic raster maps, of urban areas are important for many applications, for example mapping and navigation, urban planning and environmental monitoring, to name just a few. In spite of great progress, the task is far from solved. This is especially true for urban areas, and at high spatial resolutions (on the oder of 0.1 - 1 m). Urban land-cover classes like “road” or “building” are a mixture of many different structures and materials. As small objects such as individual cars, street furniture, roof structures, and even things like traffic signs or road markings become visible, the intra-class variability increases.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call