Skip to Main Content
 

Global Search Box

 
 
 

ETD Abstract Container

Abstract Header

Assessment of Shoreline Vegetation in the Western Basin of Lake Erie Using Airborne Hyperspectral Imagery

Abstract Details

2016, Master of Science (MS), Bowling Green State University, Geology.
Lake Erie is well known for its high biodiversity and productive fisheries. Rapid human population growth, urbanization and agricultural development cause various threats to biodiversity in the Lake Erie coastal area. Recently, there has been growing interest in fish diversity and abundance in relation to landscape, land cover and vegetation diversity. In this study, identification of land cover and assessment of biodiversity in the Lake Erie nearshore ecosystems have been conducted using a combination of remote sensing and field data. In collaboration with the University of Toledo and the Ohio Department of Natural Resources, twenty two pre-select sites along the coast of the Western basin were assessed using the airborne NASA Glenn Hyperspectral Imager (HSI) and in-situ hyperspectral measurements for mapping land cover types at different compositional levels. This study also present an insight into the different atmospheric correction models and classification techniques and their applicability to the NASA Glenn HSI data. Eight different atmospheric correction methods and ten different image classification methods were evaluated. The Empirical Line Calibration was chosen as the best atmospheric method for the NASA Glenn HSI images. The MODRAN based models such as ENVI FLAASH and ATCOR are not applicable to the NASA Glenn HSI images. The Support Vector Machine (SVM) classification method results in the highest overall accuracy (85.58%). The error propagation due to the inclusion and exclusion of NIR bands at different pre-processing and processing levels suggests that the highest accuracy was obtained when NIR bands are excluded before the atmospheric corrections (85.58%). The accuracy is somewhat lower (84.89%) when NIR bands are excluded prior to the classification, leaving the lowest accuracy for the case when NIR bands are included in both atmospheric correction and classification steps. To avoid the need of excluding NIR bands, some additional radiometric adjustment of this spectral region should be considered by NASA. The multispectral images collected by Pleiades exhibit lower classification accuracy when compared with the NASA Glenn HSI data (81.35% and 93.28% for a chosen image respectively) even in the case when the NIR bands are excluded from the hyperspectral dataset. It is most likely that spectral resolution causes the trend. The diversity indices (Shannon index and Simpson index) calculated from the NASA Glen HSI images suggest that the diversity in the sites are high (0.67 for Simpson and 1.44 for Shannon). The method is proposed as the baseline for future studies where the biodiversity indices are generated from remote sensing. A spectral library was created for tree and shrub species using the in-situ reflectance measurements to support our analysis and future studies.
Anita Simic, Dr. (Advisor)
Enrique Gomezdelcampo, Dr. (Committee Member)
Peter Gorsevski, Dr. (Committee Member)
150 p.

Recommended Citations

Citations

  • Rupasinghe, P. A. (2016). Assessment of Shoreline Vegetation in the Western Basin of Lake Erie Using Airborne Hyperspectral Imagery [Master's thesis, Bowling Green State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1467323545

    APA Style (7th edition)

  • Rupasinghe, Prabha. Assessment of Shoreline Vegetation in the Western Basin of Lake Erie Using Airborne Hyperspectral Imagery. 2016. Bowling Green State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1467323545.

    MLA Style (8th edition)

  • Rupasinghe, Prabha. "Assessment of Shoreline Vegetation in the Western Basin of Lake Erie Using Airborne Hyperspectral Imagery." Master's thesis, Bowling Green State University, 2016. http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1467323545

    Chicago Manual of Style (17th edition)