Assessment of asbestos-cement roof distribution and prioritized intervention approaches through hyperspectral imaging

Our plain-English summary of Assessment of asbestos-cement roof distribution and prioritized intervention approaches through hyperspectral imaging, by D. Enrique Valdelamar Martínez, M. Saba, L.K. Torres Gil (Heliyon, 2024).

The study

• Authors: D. Enrique Valdelamar Martínez, M. Saba, L.K. Torres Gil
• Journal: Heliyon
• Published: 2024;10(3):e25612
• Cited by: 7 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

The discernment of asbestos-cement (AC) roofs within urban areas stands as a pivotal concern pertinent to communal well-being and ecological oversight, particularly in emerging nations where asbestos continues to be extensively employed. Conventional methodologies entailing the recognition of asbestos-cement roofs and the characterization of their degradation status, such as tangible examinations and laboratory assays, prove to be temporally protracted, financially demanding, and arduous to extrapolate comprehensively across expansive urban domains. In this paper, it is presented a novel approach for identifying asbestos-cement roofs in urban areas using hyperspectral airborne acquisition and carry out a diagnosis that allows to identify the state of asbestos-cement roofs and thus provide a tool for the competent authorities to develop and prioritize intervention strategies to mitigate the problem. Four different methodologies were implemented and compared, three of which are new in the literature, to identify the deterioration of asbestos-cement (AC) roof state in large urban areas. This, in turn, furnishes a tool for competent authorities to identify the state of AC roofs, develop and prioritize intervention strategies to mitigate the problem. The control points in field allowed validating the classification and the proposed methodology for the prioritization of intervention in AC roofs. Some neighborhoods in the city showed peaks in the area of asbestos-cement roofs of 47% of the total area of the neighborhood, representing practically all of the roofs present in the neighborhood. On average around 20% of the total area of a neighborhood in Cartagena is covered by AC. Furthermore, it was found a total area of AC roofs throughout the city of more than 9 km2 (9 million square meters). On the other hand, two of the 4 methods used showed encouraging results that demonstrate their ability to identify covers in poor and good condition at a large scale from hyperspectral images. This academic novelty suggests that there is a possibility of practical application of these methods in other urban contexts with high concentrations of AC roofs, helping in the planning and optimization of intervention strategies to mitigate the risk in public and environmental health due to the presence of asbestos.

© 2024 The Authors.

In plain English

This study used aircraft-mounted hyperspectral cameras to map asbestos-cement roofs in Cartagena de Indias, Colombia, and to test ways of judging their condition from the images. Microscope analysis of 215 roof samples found chrysotile in 188, crocidolite as well in 52, amosite in one, and no asbestos in 27. Checked against field data, the roof classification had an overall accuracy of 96%. The authors identified 9.1 square kilometres of asbestos-cement roofing, 11.2% of the study area. The 11.4% of the area lying in the airport approach zone could not be surveyed. In field inspections, 125 of 149 asbestos-cement roofs (83.9%) were rated high intervention priority. The authors tested four image-based condition methods. An earlier published index reached precisions of no more than about 40%. Two new methods reached around 60%, rising to up to 70% with added synthetic data. In the imaged area, one new method classed 50.46% of asbestos-cement roofs as high priority. The authors concluded that more than half of the city's asbestos-cement roofs were in critical condition, and acknowledged that errors remained significant.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Remote-sensing study: airborne hyperspectral imaging of city roofs, combined with field roof sampling, polarized light microscopy of samples, laboratory hyperspectral tests, and comparison of four image-based methods for rating roof condition
• Who was studied: Roofs in the urban area of Cartagena de Indias, Colombia (about 1 million people, 80.9 square kilometres, 190 neighbourhoods). The authors randomly collected 215 roof samples in a field campaign. Survey dates are not given; the aerial survey took three days.
• Exposure: Asbestos-cement roofing. The asbestos found was chrysotile, with crocidolite in some samples and amosite in one. Roofs were rated by field visual inspection as high or low intervention priority.
• Main results:
  • Of 215 roof samples, 188 contained chrysotile, with a volumetric ratio relative to cement of 38.73 plus or minus 8.10%.
  • In 52 of the chrysotile-positive samples, crocidolite was also present.
  • One sample also contained amosite.
  • The remaining 27 samples were asbestos-free fiber cement.
  • Roof classification from the images had an overall accuracy of 96%.
  • According to the text, 146 of 149 asbestos-cement roofs in the surveyed area were correctly classified (producer accuracy 98%).
  • A total of 9.1 square kilometres of asbestos-cement roofs was identified in Cartagena, 11.2% of the study area.
  • Across the 190 neighbourhoods, asbestos-cement roofs covered on average 20.4% of neighbourhood area. Some neighbourhoods peaked at more than 47.6%.
  • Field assessment of the 149 asbestos-cement roofs rated 24 (16.1%) as low intervention priority and 125 (83.9%) as high intervention priority.
  • Applied to Cartagena, the threshold of the previously published moss and lichen index gave errors of up to 96% in predicting a roof's field-rated priority. At its best value for this city, its precision did not exceed 40%.
  • The two new chrysotile-reflectance-based methods (ISDCAP and R2327) had a precision of around 60%, rising up to 70% with synthetic low-priority data added.
  • Applied to the imaged study area, the R2327 method classed 49.54% of asbestos-cement roofs as low intervention priority and 50.46% as high intervention priority. This is an image-based classification, and the study area had an 11.4% data gap.
  • In a comparison of laboratory and airborne hyperspectral data on the same roofs, the method taken from the literature had the highest error (75%).
• The authors' conclusion: The authors concluded that hyperspectral identification of asbestos-cement roofs was reliable. They found that their new methods fitted the case study better than the published index, and that more than 50% of the city's asbestos-cement roofs were in critical condition (high intervention priority).
• Limitations the authors state:
  • No hyperspectral data could be collected over the central zone near the international airport, 11.4% of the study area.
  • Shadows from trees caused misclassified pixels.
  • Painted asbestos roofs were difficult to identify accurately.
  • Errors remained significant even with the two best methods. The authors attributed this partly to comparing visual field judgments with image data that do not capture small cracks, edge condition or fiber release.
  • Crocidolite and amosite mixed with chrysotile in some samples have different spectral signatures, which may reduce accuracy.
  • The image pixel size (0.8 m) differs greatly from the laboratory scale, which may influence reflectance values.
  • Shaded roof sections, typically on steeper slopes, tended to be classified as being in good condition.

Topics

Medical terms in this study, explained in our Lexicon: Asbestos-Cement.

Related studies

• Risk assessment of the decay of asbestos cement roofs — A. Campopiano, D. Ramires, et al. (2009)
• Mapping roof coverings of asbestos-cement, the first step to control the technical condition/threat and establish priorities for replacement in developing countries — L.K.T. Gil, D. Valdelamar Martínez, et al. (2024)
• Asbestos containing materials detection and classification by the use of hyperspectral imaging — G. Bonifazi, G. Capobianco, et al. (2018)
• Sample preparation method for visualization of nanoparticulate captured on mixed cellulose ester filter media by enhanced darkfield microscopy and hyperspectral imaging — N.M. Neu-Baker, A.C. Eastlake, et al. (2019)
• Mapping asbestos-cement roofing with hyperspectral remote sensing over a large mountain region of the Italian Western Alps — F. Frassy, G. Candiani, et al. (2014)
• Effect of heat stress amelioration through open-ridge ventilated thatched roof housing on production and reproduction performance of crossbred Jersey cows — D.K. Mandal, A. Mandal, et al. (2021)

Browse more research in Heliyon, or return to medical research.