In recent years, rooftops have been used as productive spaces for local resource (e.g. food, water, or energy) production. The main objective of this study is to demonstrate the potential of hyperspectral data to identify roof materials in cities. The case study area is located in a Mediterranean region north of Barcelona (Spain); diverse building typologies, building uses (residential, industrial, office, services and retail) and population densities constitute this study area. Two hyperspectral remote sensors (HRSs) were used to obtain hyperspectral images. Seven roof classes were identified using the k-means clustering algorithm. Most roofs were identified as metals (33 percent), followed by ceramic tiles (22 percent). The findings showed that 51 percent of roofs could be suitable for food, water, or energy production. Nevertheless, for a more reliable assessment of this potential, more criteria are needed.