A comparison of object-oriented and pixel-based classification methods for mapping land cover in northern Australia

The development of robust object-oriented classification methods suitable for medium to high resolution satellite imagery provides a valid alternative to ‘traditional’ pixel-based methods. This paper compares the results of an object-oriented classification to a supervised pixel-based classification for mapping land cover in the tropical north of the Northern Territory. The object-oriented approach involved the segmentation of image data into objects at multiple scale levels. Objects were assigned class rules using spectral signatures, shape and contextual relationships. The rules were then used as a basis for the fuzzy classification of the imagery. The supervised pixel-based classification involved the selection of training areas and a classification using maximum likelihood algorithm. Accuracy assessment of bothe classifications were undertaken. A comparison of the results shows better overall accuracy of the object-oriented classification over the pixel-based classification. This object-oriented method provided results with accepotable accuracy; indicating object-oriented analysis has great potential for extracting land cover information from satellite imagery captured over tropical Australia.