Spatial Assessment of Asbestos Mine Remediation Effect Using Airborne Hyperspectral Imaging System

This work investigated an abandoned asbestos mine (AAM) remediation project in CA, US. Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) hyperspectral data were used to assess the mine treatment effect. The distribution of AAM and restored area were evaluated before and after remediating based on the spectral analysis and model for indicating naturally occurring asbestos (NOA) and encapsulation (remediation). We developed NOA, host rock, and encapsulation indices by binary logistic regression modeling using hyperspectral bands. The detection models statistically selected visible-near infrared (VNIR) regions rather than shortwave infrared (SWIR) ranges. The models-based classification accuracy was approximately 84%. Notably, the detection accuracy of non-treated and treated areas was increased to about 90% excluding the host rock index. The NOA and encapsulation indices showed that they can be efficiently applied to asbestos occurrence/remediation. The remote sensing data revealed that the whole AAM area was increased by ~5% by the remediation process, and the restoration activity reduced asbestos exposure by ~32%. This work would be contributed to providing an intuitive and realistic-spatial understanding of the planning and managing remediation project.