Mapping Asbestos-cement Roofs Using Remote Sensing

Italy’s Ministry of the Environment and Energy Security has assigned the project to a team headed by e-GEOS of the Leonardo Group, working alongside MapSat, Planetek Italia, and the University of Cassino.

Satellite Detection Initiative

Asbestos Cement Inventory: High-Resolution Analysis

The program will update the register of asbestos-cement rooftops on public and private properties throughout Italy through an ongoing, phased workflow that is refined as new imagery is acquired and verification results are incorporated.Remote sensing data. The workflow can combine spectral signature analysis and material-focused indices with object-based image analysis (segmenting roofs into meaningful objects rather than classifying pixel by pixel). Ancillary layers such as building footprints and land-use masks can be used to focus the search on roof surfaces and reduce confusion from roads, bare soil, and other non-roof targets.Multispectral satellite imagery (very high resolution). Very-high-resolution commercial scenes are used where individual rooftops must be separated and characterized; common inputs include sub-meter panchromatic imagery paired with meter-class multispectral data. When available, shortwave-infrared coverage is particularly useful for separating roofing materials with similar visible-color appearance, while visible and near-infrared bands support material discrimination and shadow handling.Artificial intelligence-driven classification. Models such as Random Forest and Support Vector Machines are often used for tabular spectral-and-texture features, while convolutional neural networks can be applied to image chips or segmented roof objects for pattern recognition. Performance is typically reported with confusion-matrix metrics (overall accuracy plus class-level precision and recall), and validation relies on independent reference samples (for example, municipal records, targeted inspections, and photo interpretation), with error tracked as false positives and false negatives.

Study areas used to calibrate and check satellite mapping are typically selected to represent different building styles and roofing stocks, including dense urban cores, industrial zones, and smaller municipalities. In Italy, representative verification areas commonly include provinces and metropolitan areas such as Milan and Brescia (Lombardy), Bologna (Emilia-Romagna), Padua (Veneto), Rome and Frosinone/Cassino (Lazio), Naples (Campania), and Palermo (Sicily), alongside smaller towns where roof types and weathering patterns differ.

Topography can materially affect detection. In hilly and mountainous terrain, slope and aspect change illumination and viewing geometry, increasing shadowing and altering apparent reflectance; this can reduce separability between asbestos-cement roofs and visually similar materials. Elevation-driven differences in haze, seasonal snow, and vegetation phenology can further complicate classification, so accuracy is often higher in flat, uniformly lit urban areas than in steep valleys and ridge-settlements unless terrain-aware correction and careful sampling are applied.

Costs vary significantly by technology and by the level of verification required. Satellite-based approaches generally have lower marginal cost for broad, repeated national coverage, but their total cost depends on the spatial resolution purchased, cloud-free tasking windows, and the amount of processing and quality control needed. Aerial surveys can deliver finer detail and more flexible acquisition timing over limited areas, but typically carry higher acquisition costs per square kilometer. Ground-based mapping provides the most direct confirmation, yet is labor-intensive and is usually reserved for targeted follow-up, prioritization, and enforcement where remote sensing indicates elevated likelihood or uncertainty.

The public-health relevance is direct: asbestos-cement products can release hazardous fibers as they weather, break, or are disturbed during maintenance and demolition, and inhalation exposure is associated with serious diseases such as asbestosis, lung cancer, and mesothelioma. Mapping supports prevention by helping authorities prioritize inspections, communicate risk, plan remediation, and track progress over time—especially where building density makes manual inventorying impractical.

Key limitations of satellite detection include spectral confusion with non-asbestos roofing (for example, certain fiber-cement, concrete, or weathered surfaces), mixed pixels on small or complex roofs, and occlusion from trees, rooftop equipment, and deep shadows. Cloud cover and seasonal lighting differences can reduce usable acquisitions, and roof aging or coatings can shift the material signature, increasing both missed detections and false alarms; for this reason, satellite results are typically treated as a decision-support layer that benefits from targeted on-the-ground confirmation.Accurate asbestos-cement maps turn a widespread hazard into a prioritized, trackable remediation workload, helping public agencies focus inspections and interventions where they reduce exposure risk the most.“Geoinformation proves its worth when it protects communities and directly supports public institutions, ultimately improving citizens’ health,” said C. Milena A. Lerario, chief executive officer of e-GEOS. “Being part of Leonardo enables us to combine deep expertise, advanced technologies, and system integration with very high-resolution satellite imagery and artificial intelligence-powered analytics. This capability delivers a broader and more detailed picture of the territory, helping address complex issues and underpinning evidence-based policies with timely, accurate data to safeguard the environment and enhance quality of life.”