A comparison of supervised and rule-based object-orientated classification for forest mapping

[摘要] ENGLISH ABSTRACT: Supervised classifiers are the most popular approach for image classification due to their highaccuracies, ease of use and strong theoretical grounding. Their primary disadvantage is the highlevel of user input required during the creation of the data needed to train the classifier. Onealternative to supervised classification is an expert-system rule-based approach where expertknowledge is used to create a set of rules which can be applied to multiple images. This researchcompared supervised and expert-system rule-based approaches for forest mapping. For thispurpose two SPOT 5 images were acquired and atmospherically corrected. Field visits, aerialphotography, high resolution imagery and expert forestry knowledge were used for thecompilation of the training data and the development of a rule-set. Both approaches wereevaluated in an object-orientated environment. It was found that the accuracy of the resulting mapswas equivalent, with both techniques returning an overall classification accuracy of 90%. Thissuggests that cost-effectiveness is the decisive factor for determining which method is superior.Although the development of the rule-set was time-consuming and challenging, it did not requireany training data. In contrast, the supervised approach required a large number of training areasfor each image classified, which was time-consuming and costly. Significantly more training areaswill be required when the technique is applied to large areas, especially when multiple images areused. It was concluded that the rule-set is more cost-effective when applied at regional scale, but itis not viable for mapping small areas.

[发布机构] Stellenbosch University