Buildings Classification using Very High Resolution Satellite Imagery

Abstract

Buildings classification using satellite images is becoming more important for several applications such as damage assessment, resource allocation, and population estimation. We focus, in this work, on buildings damage assessment (BDA) and buildings type classification (BTC) of residential and non-residential buildings. We propose to rely solely on RGB satellite images and follow a 2-stage deep learning-based approach, where first, buildings' footprints are extracted using a semantic segmentation model, followed by classification of the cropped images. Due to the lack of an appropriate dataset for the residential/non-residential building classification, we introduce a new dataset of high-resolution satellite images. We conduct extensive experiments to select the best hyper-parameters, model architecture, and training paradigm, and we propose a new transfer learning-based approach that outperforms classical methods. Finally, we validate the proposed approach on two applications showing excellent accuracy and F1-score metrics.

Classification of Building Types in Germany: A Data-Driven Modeling Approach

“…In addition to these features, the building footprints are gathered using edge-based geometric grouping or object-based classification [6]. The labeling of buildings or the classification of building types were accomplished in [5,[7][8][9][10][11][12][13][14][15][16][17] using the datasets produced from the aforementioned approaches using remote sensed images, google earth images [18,19] and LiDAR data. Ref.…”

“…However, because the experts' high degree of semantics highlights a semantic gap in the remote sensed imaging data [7], Point-of-Interest (POI) data (i.e., specific point location, e.g., hospital, office, restaurant, etc. ), which is often user-generated data, was mapped to the remote sensed data to identify building types [10,11]. In this context, ref.…”