International Journal of Geographical Information Science, Volume 25, Issue 6, 2011
Javier Leon and Colin D. Woodroffe
“Monitoring coral reefs is of great importance for environmental management of these ecosystems. The use of remote sensing and geographical information systems enables rapid and effective mapping of the geomorphology of reefs that can be used as a basis for biodiversity and habitat assessments. However, pixel-based approaches have not been appropriate for detailed mapping of such complex systems. An object-based image analysis (OBIA) approach was used in this study to map intra-reef geomorphology of coral reefs across the Torres Strait region using Landsat ETM+ imagery. By combining image analysis techniques and a non-parametric neural network classifier and incorporating additional spatial information such as context, shape and texture, the accuracy of the segmentation and classification was improved considerably.
Qualitative assessment of segmentation. (a) Segmentation based on Landsat image (30 m) (red outlines) and (b) pan-sharpened image (15 m) (blue outlines). Arrows show examples of under-segmentation on the coarser-resolution image when compared to the reference data derived from a Quickbird image (black outlines) over the western end of Bet reef platform. Clouds and shadows appear masked (stippled areas).
“A large-scale synoptic map of 10 geomorphological classes was produced for Torres Strait with an overall accuracy of 75%. The OBIA approach employed in this research has enabled the geomorphology of reef platforms to be mapped for the first time at such accuracy and descriptive resolution.”
International Journal of Applied Earth Observation and Geoinformation, Volume 13, Issue 6, December 2011, Pages 884-893
Timothy G. Whiteside, Guy S. Boggs, and Stefan W. Maier
- Per-pixel classification of high spatial resolution imagery can produce spurious pixels.
- Maximum Likelihood Classifier compared to object-based supervised classification.
- Object-based classification statistically superior to per-pixel approach.
- Object-based classification suited to heterogeneous landscapes such as savanna.
“The development of robust object-based 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-based classification to a supervised per-pixel classification for mapping land cover in the tropical north of the Northern Territory of Australia. The object-based approach involved segmentation of image data into objects at multiple scale levels. Objects were assigned classes using training objects and the Nearest Neighbour supervised and fuzzy classification algorithm. The supervised pixel-based classification involved the selection of training areas and a classification using the maximum likelihood classifier algorithm. Site-specific accuracy assessment using confusion matrices of both classifications were undertaken based on 256 reference sites. A comparison of the results shows a statistically significant higher overall accuracy of the object-based classification over the pixel-based classification. The incorporation of a digital elevation model (DEM) layer and associated class rules into the object-based classification produced slightly higher accuracies overall and for certain classes; however this was not statistically significant over the object-based using spectral information solely. The results indicate object-based analysis has good potential for extracting land cover information from satellite imagery captured over spatially heterogeneous land covers of tropical Australia.”