The purpose of the present thesis is to show that hyperspectral data can be used for discrimination between different tree species. The data set used in this study contains the hyperspectral measurements of leaves of seven savannah tree species. The data is high-dimensional and shows large within-class variability combined with small between-class variability which makes discrimination between the classes challenging. We employ two classification methods: G-nearest neighbour and feed-forward neural networks. For both methods, direct 7-class prediction results in high misclassification rates. However, binary classification works better. We constructed binary classifiers for all possible binary classification problems and combine them with Error Correcting Output Codes. We show especially that the use of 1-nearest neighbour binary classifiers results in no improvement compared to a direct 1-nearest neighbour 7-class predictor. In contrast to this negative result, the use of neural networks binary classifiers improves accuracy by 10% compared to a direct neural networks 7-class predictor, and error rates become acceptable. This can be further improved by choosing only suitable binary classifiers for combination.

M.Sc. (Botany) / Global efforts to protect the world’s forests from unsustainable and inequitable exploitation have been undermined in recent years by rampant illegal logging in many timber-producing countries. A prerequisite for efficient control and seizure of illegally harvested forest product is a rapid, accurate and tamper proof method of species identification. DNA barcoding is one such a tool, relatively simple to apply. It is acknowledged to bring about accuracy and efficiency in species identification. In this study a DNA barcode reference library for traded and protected tree species of southern Africa was developed comprising of 81 species and 48 genera. Four primary analyses were conducted to assess the suitability of the core barcodes as a species identification tool using the R package Spider 1.2-0. Lastly, to evaluate this identification tool, query specimens independently sampled at a Mozambican logging concession were identified using DNA barcoding techniques. The nearest neighbour (k-NN) and best close match (BCM) distance based parameter yielded 90% and 85% identification success rate using the core plant barcodes respectively. DNA barcoding identification of query specimens maintained a constant 83% accuracy over the single marker dataset and the combined dataset. This database can serve as a backbone to a control mechanism based on DNA techniques for species identification and also advance the ability of relevant authorities to rapidly identify species of timber at entry and exit points between countries with simple, fast, and accurate DNA techniques.