Urban planning applications (energy audits, investment, etc.) require anunderstanding of built infrastructure and its environment, i.e., bothlow-level, physical features (amount of vegetation, building area and geometryetc.), as well as higher-level concepts such as land use classes (which encodeexpert understanding of socio-economic end uses). This kind of data isexpensive and labor-intensive to obtain, which limits its availability(particularly in developing countries). We analyze patterns in land use inurban neighborhoods using large-scale satellite imagery data (which isavailable worldwide from third-party providers) and state-of-the-art computervision techniques based on deep convolutional neural networks. For supervision,given the limited availability of standard benchmarks for remote-sensing data,we obtain ground truth land use class labels carefully sampled from open-sourcesurveys, in particular the Urban Atlas land classification dataset of $20$ landuse classes across $~300$ European cities. We use this data to train andcompare deep architectures which have recently shown good performance onstandard computer vision tasks (image classification and segmentation),including on geospatial data. Furthermore, we show that the deeprepresentations extracted from satellite imagery of urban environments can beused to compare neighborhoods across several cities. We make our datasetavailable for other machine learning researchers to use for remote-sensingapplications.
展开▼
