Remote sensing for assessing vegetated roofs with a new replicable method in Paris, France

Vegetated roofs provide many ecosystem services and support urban biodiversity. While it would be interesting to study the contribution of vegetated ecological corridors, are listed in no French databases. Because their intrinsic nature as roofs, small number, size, type vegetation planted on them, seem very difficult identify. We propose a method automatically identify roofs. Using infrared aerial photographs building shape data, we were able build model detecting using remote sensing supervised classification techniques. The major difficulty lies distinguishing between real partially covered by tree foliage growing ground. In this operation, our obtains an error rate ∼18 % . improve knowledge detection cities. Moreover, opens perspectives analysis networks cities function height. addition, could tool for municipalities monitor development prioritize planning.