robosat

worked on before: * https://github.com/Project-OSRM/osrm-backend * https://github.com/mapbox/robosat Bonus points if you have worked professionally with geospatial data and software before and/or have contributed to OpenStreetMap, be it code or map edits. Write me a note at daniel(dot)hofmann(at)daimler(dot)com telling me about what projects you've worked on that fit what we're looking for. Include a robot emoji in the subject to confirm you are not a robot ;) You can also find us and talk to us engineers at local meetups such as Geo Berlin.

esources are going into creating and cleaning datasets, doing hard-negative mining, running multiple training iterations and so on. They're also bound to the Mapbox aerial imagery on specific zoom levels. The model architecture is kept simple on purpose. It used to be an encoder-decoder U-Net'ish architecture which we trained from scratch. Recently (https://github.com/mapbox/robosat/pull/46) I switched out the encoder to a pre-trained ResNet, as proposed by Alexander Buslaev. It's a mix of the papers listed in the docstring at the top with a focus on simplicity and maintainability: https://github.com/mapbox/robosat/blob/1e687552fe9b254a14d55... Internally we were also exploring a multi-class PSPNet but decided not to move forward with it right now: the RoboSat model is currently a binary model (feature vs. background) which makes a few things easier in practice, such as efficiently storing results which is needed when scaling it up e.g. to all of North America.