Application of Landsat data for land use classification and Wetland study in West Tennessee
The Obion-Forked Deer River Basin in northwest Tennessee is confronted with several acute land use problems which contribute to excessive erosion, sedimentation, pollution, and hydrologic runoff. In order to assist with resource planning in this basin, the objective of this study was to demonstrate the application and utilization of Landsat data in imagery form for determining land use of selected watershed areas within the basin, with special emphasis on determining wetland boundaries.
Densitometric analysis was performed to allow the numerical classification of objects observed in the imagery on the basis of measurements of optical densities. Relative light transmission measurements were taken on four types of scene elements in each of three Landsat black and white bands in order to determine which classifications could be distinguished. The analysis of band 6 determined forest and agricultural classifications to be significantly different, but the urban and wetlands classifications were not. Both bands 4 and 5 showed a significant difference existed between the combined classification of wetlands-agriculture and urban areas. Therefore, the combination of band 6 with either band 4 or 5 would permit the separation of the urban from the wetland classification.
In order to enhance the urban areas as well as the wetland boundaries, the Landsat black and white bands were combined in a multispectral additive color viewer. Several combinations of filters and light intensities were used to obtain maximum discrimination between points of interest. The best results for enhancing wetland boundaries and urban areas were achieved by using a color composite (i.e. a blue, green, and red filter on bands 4, 5, and 6 respectively).
Using the Tennessee Land Use Classification System, Level I classifications were determined on a 1:250,000 scale Landsat color composite and on 1:130,000 high altitude color infrared photography for a specified watershed. Level II classifications were done on 1:24,000 scale photography. Areas were calculated for the various classifications at each scale. The percentage of difference or error was calculated between the low altitude control data and each of the other scales in order to indicate their level of accuracy.
Differences in acreage between the Landsat imagery and the low altitude imagery related primarily to resolution limitations of the Landsat data and not to an inability to differentiate between classifications. Use of the additive color viewer demonstrated that the desired classifications could be separated. From this study, it was concluded that Landsat data in imagery form can be used effectively to determine wetland boundaries, and can therefore be used to periodically measure remaining acreages and types of wetlands, thereby providing needed resource information for management and preservation programs.
Thesis77.J662.pdf
11.18 MB
Adobe PDF
e2e81bd99ae078c724a1a7b371573ce0