- Title
- Spectroscopy based mapping of selected dominant rock types in the Balfour Formation (Beaufort Group), Eastern Cape Province, South Africa
- Creator
- Qabaqaba, Mcebisi
- Subject
- spectrum Analysis
- Date
- 2019
- Type
- Thesis
- Type
- Masters
- Type
- MSc
- Identifier
- http://hdl.handle.net/10353/14693
- Identifier
- vital:40057
- Description
- Lithological mapping of areas is crucial for both mineral resource identification and determination of ore concentrations. Without adequate knowledge on the extent of an area’s lithology, the above applications cannot be meaningfully accomplished. Spectroscopy-based mapping of rocks involves the use of portable reflectance spectrometers and digital maps to analyze, assess and map changes in rocks. Thereafter, understand and interpret the geochemical, physical and environmental processes that have an impact on the reflectance of the rocks while also making spatial and spectral links. Mapping lithology using remote sensing involves identification of rocks in remotely sensed images with lithological mapping being achievable through field and laboratory spectroscopy and creation of a spectral library. The main goal of this study was to measure, extract and analyze spectral signatures of dominant rocks within the Balfour Formation in order to create a spectral library that is capable of enabling, fast and cost-effective mapping of the spatial distributions of rocks within the Balfour Formation. The spectra of dominant rocks of the Balfour Formation were measured using the SR 3500 spectroradiometer. Different pre-processing techniques were applied to the acquired spectra such as smoothing, scatter correction and outlier detection. Random Forest classifier was used in the classification of the different rocks. Analysis of Variance and Bonferroni post hoc test were applied on the spectral signatures of different rocks to ascertain if there were any spectral differences between the absorption features in the spectra of the rocks. Landsat 8 imagery was used for classification of the different rocks within the Balfour Formation, maximum likelihood and spectral angle mapper were the two classifiers that were XI used. Maximum likelihood presented an accuracy of 76.19 % and 66.67 % for spectral angle mapper.
- Format
- 187 leaves
- Format
- Publisher
- University of Fort Hare
- Publisher
- Faculty of Science and Agriculture
- Language
- English
- Rights
- University of Fort Hare
- Hits: 533
- Visitors: 572
- Downloads: 63
Thumbnail | File | Description | Size | Format | |||
---|---|---|---|---|---|---|---|
View Details | SOURCE1 | Qabaqaba Mcebisi MSc Final Version.pdf | 2 MB | Adobe Acrobat PDF | View Details |