Automatic remote sensing of asbestos-cement roofs
A method for the automatic recognition and subsequent mapping of visible surfaces characterized by the significant presence of asbestos fibers and/or mixtures of cement and asbestos (for example, asbestos roofs). The technology allows with precision, simplicity, reliability, and efficiency to identify areas with significant presence of asbestos in images acquired from satellite or aircraft, on the basis of an image comparison that overcomes any problems of inhomogeneity that characterize other image acquisition processes.
The current methods of synoptic mapping (remote detection) are based on the use of a supervised and subjective technique and require the use of an unsustainable amount of resources, without any guarantee of homogeneity of the result in the face of an error percentage close to the present technology. The use of an unsupervised technique is useful for the exploitation of massively multichannel hyperspectral aerial data which cannot be analyzed in supervised mode. The method which operates in unsupervised mode performs a multispectral measurement of the areas characterized by the presence of asbestos-cement or asbestos acquiring images by passive multi-spectral radiometers. This approach allows the use of electronic equipment, associated with the radiometer that acquires data either in flight or on the ground. The technology has also proven to be effective on technically advanced data (digital colour orthophotos) but not particularly sophisticated. The success rate in detention and perimeter is likely to increase in the case of the use of more sophisticated hypersprectral radiometers on 0.6-2.3 micron wavelengths. Current TRL is 9.
- Mapping of large urban and / or industrial areas for the identification of asbestos-cement roofs for their subsequent remediation.
- Automatic unsupervised pre-analysis of large areas in a very short time
- The exclusive automatic procedure guarantees the uniformity and objectivity of the analysis;
- Reduction of costs and detection times, each by at least 50% compared to photointerpretation;
- It significantly improves detection stability and mapping effectiveness.