An automatic data driven approach to derive photovoltaic-suitable roof surfaces from ALS data

Nowadays renewable energy systems (RES) become mandatory due to the fact that fossil energy declines. Solar energy is one of the key technologies in the discussion about providing urban areas with energy. Two concepts face each other, centralized and decentralized energy systems. To avoid additional land consumption with large photovoltaic (PV) installations and controversial power lines across the country as in the case of a centralized energy plant it is essential for the industry and several stakeholders to know where suitable areas for PV installations exist or could be located to build up an efficient decentralized energy network. In this paper we present an automatic approach to extract suitable single roof planes from airborne laser scanning data. The major advantage of this approach is that not only solar radiation values are calculated or ratings for buildings made, but also each single roof is extracted and presented to the different customers. Each planar roof area will have exact geometrical attributes like size, exposition and slope as well as energy yield values and approximate values about the return of investment. Shadowing effects from surrounding objects are also taken into account to detect PV suitable roof models.