Description: Microsoft says the building extraction was done in two stages:1.Semantic Segmentation – Recognizing building pixels on the aerial image using DNNs2.Polygonization – Converting building pixel blobs into polygonsWe developed a method that approximates the prediction pixels into polygons making decisions based on the whole prediction feature space. This is very different from standard approaches, e.g. Douglas-Peucker algorithm, which are greedy in nature. The method tries to impose some of a priory building properties, which are, at the moment, manually defined and automatically tuned. Some of these a priori properties are:1.The building edge must be of at least some length, both relative and absolute, e.g. 3 meters2.Consecutive edge angles are likely to be 90 degrees3.Consecutive angles cannot be very sharp, smaller by some auto-tuned threshold, e.g. 30 degrees4.Building angles likely have very few dominant angles, meaning all building edges are forming angle of (dominant angle ± nπ/2)In near future, we will be looking to deduce this automatically from existing building information. We track various metrics to measure the quality of the output. Estimated building matching metrics:Precision 99.3% Recall 93.5% Our metrics show that in the vast majority of cases the quality is at least as good as data hand digitized buildings in OpenStreetMap. It is not perfect, particularly in dense urban areas but it is still awesome.
Service Item Id: 29b889c91ec64220aacc73e2d668b7af
Copyright Text: Credit given to Microsoft and Flathead County GIS would be appreciated when deriving products from this data.