Acoustic Absorption Mapping: Wide-area Estimation of Ground Acoustic Absorption Coefficient Using Airborne Hyperspectral Imagery

Recent advances in outdoor acoustic simulation techniques and propagation modeling have enabled wide-area environmental noise assessment as notably represented by noise mapping. However, among main input information for such an assessment, ground acoustic properties are not available as a geographical dataset whereas topography is readily obtainable as a digital surface model. In this paper, a wide-area estimation technique of the ground acoustic absorption coefficient using remote sensing is explored. The selected sensor is a CASI-3 airborne hyperspectral imager, which provides high spectral and spatial resolutions in visible and near-infrared optical wavelengths. Hyperspectral images of 34 bands with a ground sampling distance of 1m were acquired for a region of approximately 3400 x 4200 square meters. Simultaneously, the ground acoustic properties were measured using an in-situ technique at 8 locations in the imaged region, each of which has 36 measurement points. A modeling process using a multiple regression analysis derived a set of equations that estimates the acoustic absorption coefficient from the 160 to the 3150Hz one-third-octave bands using the significant 3 bands of the optical spectral reflectance. The acoustic absorption map obtained by applying the derived model to the entire image qualitatively matches the known properties for various ground types.