A QGIS Tool for Automatically Identifying Asbestos Roofing

The exposure to asbestos fibers implies a long-term risk for human health, therefore the development of information systems able to detect the extent and status of asbestos over a certain territory has become a priority. This work presents a tool based on GIS Open Source software, QGIS, conceived for automatically identifying buildings with asbestos roofing. The area under investigation is the metropolitan area around Prato (I). The performance analysis of this system was carried out by classifying images obtained with the WorldView-3 sensor. These images are available at a low cost, if compared with those obtained by means of aerial surveys, and provide adequate resolution levels for roofing classification. The tool, a QGIS plugin, has shown quite good performance in identifying asbestos roofing with both some false negatives and some false positives when applying per-pixel classification. Performance improvement is obtainable when considering the percentage of asbestos pixels contained in each roof of the analyzed image. This value is also available with the plugin. In the future, this tool should make it possible to monitor over time the asbestos roof removal process in the area of interest, according to other image data entry giving evidence of such removal.