Automatic Building Extraction in Aerial Scenes Using Convolutional Networks

TL;DR

This paper presents a convolutional neural network that effectively extracts buildings from aerial images by using multi-stage activations and signed distance functions, trained on GIS data, outperforming previous methods.

Contribution

The paper introduces a novel CNN architecture with a multi-stage integration and signed distance boundary representation for improved building extraction from aerial imagery.

Findings

Achieves superior accuracy on large, complex datasets

Utilizes abundant GIS building footprint data for training

Demonstrates scalability and robustness of the method

Abstract

Automatic building extraction from aerial and satellite imagery is highly challenging due to extremely large variations of building appearances. To attack this problem, we design a convolutional network with a final stage that integrates activations from multiple preceding stages for pixel-wise prediction, and introduce the signed distance function of building boundaries as the output representation, which has an enhanced representation power. We leverage abundant building footprint data available from geographic information systems (GIS) to compile training data. The trained network achieves superior performance on datasets that are significantly larger and more complex than those used in prior work, demonstrating that the proposed method provides a promising and scalable solution for automating this labor-intensive task.