Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Characterizing scale -dependent spatial structure in multi-resolution, multi-temporal and hyperspectral remotely sensed imagery
Spatial structure information characterized from remotely sensed imagery can be used in various applications. Previous studies in this field were mostly concentrated on forests, grasslands or wetlands. Little attention was paid to the crop canopies, where this information is needed for remote sensing of agriculture. In this research, the scale-dependent spatial structure in remotely sensed imagery from different sources acquired over agricultural crop canopies was characterized to assess the effect on observed spatial structure stemming from changes in spatial, spectral and temporal factors. Three research questions were addressed: (1) What is the effect of spatial resolution on observed spatial structure and how good is the correspondence between measured and rescaled data sets? (2) How does the temporal development of spatial structure differ among spectral wavebands and indices? (3) How does observed spatial structure change with spectral resolution and wavelength? The study was conducted at the University of Nebraska Agricultural Research and Development Center, near Mead, Nebraska. Spatial structure characterization was carried out using lacunarity, semivariogram, and scale of fluctuation analyses. Results indicated that: (1) Due to the difference in spatial structure of fractional vegetation cover (FVC), one crop FVC has higher lacunarity and more complicated lacunarity deviation than the other. (2) Regularization can lead to the decrease in amount of spatial variability and increase in spatial dependence, and rescaling via block averaging can overly retain original spatial structure, which may be reduced via spatial filtering operations prior to rescaling. (3) Spatio-temporal patterns are different between vegetation indices and their component bands, and vegetation indices are more sensitive to the development of crop canopies than their component bands. (4) The degree of spatial variation with wavelength is correlated to the degree of the difference in reflectance of underlying targets and increasing bandwidth may blur spatial structure that occurs at finer spectral resolution. These findings are useful for studies of rescaling, crop monitoring, sampling design, and identification of an appropriate spatial resolution for agricultural remote sensing. ^
Chen, Weirong, "Characterizing scale -dependent spatial structure in multi-resolution, multi-temporal and hyperspectral remotely sensed imagery" (2004). ETD collection for University of Nebraska - Lincoln. AAI3149623.