Authors: P. Roy, J.O. Brumfield and A. Vaseashta
Affilation: Graduate program in Physical Sciences, United States
Pages: 631 - 634
Keywords: satellite sensors, pollution, monitoring, spectral signatures
This paper describes an approach to assess atmospheric pollutant in the troposphere and their spectral characteristic signatures by using high-spectral and spatial resolution ASTER data and its analysis of variance with in-situ data obtained using ground based sensors. The objective of this investigation is to analyze air pollutants, in urban and industrial areas, from satellite images, and to create a model with climatic data to observe the transport of air pollutant from its source, using high spectral and spatial resolution ASTER band in SWIR and TIR ranges and in-situ pollutants concentration data provided by EPA air quality monitoring stations and facilities emission data. Air quality and emission data, with spatial references, from EPA monitoring stations are used to locate sources and concentrations of pollutants using processed satellite images. ASTER data are geo-registered and geo-corrected by image-to-image registration process with the respective DOQQs. Data Fusion techniques are used to increase spatial resolution of the imagery while preserving spectral characteristics of the ASTER data. Principal Component Analysis technique is applied to transform the number of correlated bands into a smaller number of uncorrelated bands. “Density Slicing” and “Band Ratioing” techniques are also applied to extract atmospheric pollutants in the datasets. Spectra signatures of the pollutants are compared by traversing techniques in several band sets. Despite broader bandwidth of the ASTER sensor systems as compared to hyper-spectral data, we have achieved an excellent correlation with image processing techniques in all study areas under investigation. Using ‘Analysis of Variance’ from city to city, the work is in progress towards determining correlation with EPA sensors data and the ASTER pixel reflectance digital numbers, in all possible absorption bands. Furthermore, the data as obtained from processed satellite imagery is compared with our existing program of monitoring pollutants in major cities employing nanostructured materials based pollution sensors.