By Kate Lewis Hood 23 June 2022 Los Angeles, California, United States RegisTree A section of an aerial view map of trees in Los Angeles. Image source: Registree [image]. Retrieved 23 June 2022, from https://registree.ethz.ch/explore-demos.html RegisTree (a collaboration between ETH Zürich and CalTech) uses deep learning to detect and inventory street trees from publicly available images of a city. Combining geo-referenced aerial and streetview images from Google Maps with map data, the RegisTree system identifies a tree's location, species, and approximate trunk diameter, and estimates its state of health. The aim of the system is to allow local councils managing urban trees to locate areas where trees need further maintenance, and where new trees could be planted. RegisTree has so far been piloted in Los Angeles.