Mapping bugweed (Solanum mauritianum) infestations in Pinus patula plantations using hyperspectral imagery and support vector machines

[摘要] The invasive plant known as bugweed (Solanum mauritianum) is a notorious invader offorestry plantations in the eastern parts of South Africa. Not only is bugweed considered to beone of five most widespread invasive alien plant (IAP) species in the summer rainfall regionsof South Africa but it is also one of the worst invasive alien plants in Africa. It forms denseinfestations that not only impacts upon commercial forestry activities but also causessignificant ecological and environment damage within natural ecotones. Effective weedmanagement efforts therefore require new and robust approaches to accurately detect; mapand monitor weed distribution in order to mitigate the impact on forestry operations. In thisregard, support vector machines (SVM) offer a promising alternative to conventional machinelearning and pattern recognition approaches to weed detection and mapping using remotesensing. The main objective of this research was to determine the utility of using a recursivefeature elimination support vector machine (SVM-RFE) based approach with a 272-wavebandAISA Eagle image to detect and map the presence of co-occuring bugweed within maturePinus patula compartments in KwaZulu Natal. The SVM-RFE approach required only 17optimal bands from the original 272 band image to produce a classification accuracy of 93%and True Skills Statistic of 0.83. Results from this study indicate that (1) there is definitepotential for using SVMs for accurate detection and mapping of bugweed species incommercial plantations and (2) it is not necessary to use the entire 272-band dataset toaccurately detect bugweed occurrence as the SVM-RFE approach will identify an optimalsubset of wavebands for weed detection enabling substantially improved data processing andanalysis.

[发布机构] University of Pretoria