The focus of this proposal is to use Hyperspectral Imagery from airborne platforms to provide critical information for water quality monitoring. High dimensionality 224 spectral channel data the 400 to 2500um wavelength region, mainly AVIRIS, demonstrate the feasibility of using advanced remote sensing techniques to discriminate and distinguish submerging aquatic vegetation (chlorophyll a), and suspended sediments concentration and other water quality parameters. These hyperspectral data can assist us to develop a better understanding of light/water/substance interactions. Also, correlation between hyperspectral spectral index and the field sampling measurements in Chesapeake Bay is essential to identify validation patterns. Such information should allow us to move away from empirical approaches now being used. That will overcome the limitation of conventional multi-spectral techniques for measuring water quality involved in situ measurements and the collection of water samples for subsequent laboratory analyses. The HSI advanced techniques will allow us to use the full resolution electromagnetic spectrum to monitor water quality parameters that can develop new models and algorithms. Our goal is to determine how hyperspectral imaging might benefit those currently involved in monitoring water quality, based on the results of field study and similar studies in other application areas. Also, the study includes advanced techniques for image analysis applies to both small and large oil spill targets using spectral analysis. Our study is focusing on using spectral analysis for oil spill detection techniques. These different techniques will lead to accurate identification of surface material types and extraction of their spectra from hyperspectral imagery.

The Geography Seminar Series is organized by the Geographical Honors Society Gamma Theta Upsilon Eta Omicron Chapter.

The Landsat satellites have been providing earth observation data continuously since early 1970s, and form a cornerstone for mid-resolution remote sensing. Continuity of Landsat-like data is a critical need within the Earth Science community. However, the failure of the Scan-Line Corrector (SLC) mechanism on Landsat 7 in 2003 and increasing age of Landsat 5 have threatened this continuity. While a new Landsat Data Continuity Mission (LDCM) satellite will begin operation in 2011, the difficulties in maintaining Landsat continuity have highlighted the need to combine the capabilities of existing international sensors to provide a more robust observational record. Combining different mid-resolution sensors can also provide more frequent observations throughout the growing season for monitoring rapid vegetation phenological changes. In this presentation, I will present our recently developed general empirical relation model (GERM) that uses MODIS surface reflectance as reference data set to normalize multiple mid-resolution sensor data to a consistent data stream. I will also present the Spatial and Temporal Adaptive Reflectance Fusion Model (StarFM) algorithm to blend Landsat and MODIS surface reflectance. Using this approach, high-frequency temporal information from MODIS and high-resolution spatial information from Landsat can be blended for applications which require high resolution in both time and space. The MODIS daily 500m surface reflectance and the 16-day repeat cycle Landsat ETM+ 30m surface reflectance are used to produce a synthetic “daily” surface reflectance product at ETM+ spatial resolution.