Examining the Effect of Ancillary and Derived Geographical Data on Improvement of Per-Pixel Classification Accuracy of Different Landscapes
- 155 Downloads
Effective conservation and management of natural resources requires up-to-date information of the land cover (LC) types and their dynamics. Multi-resolution remote sensing (RS) data coupled with additional ancillary topographical layers (both remotely acquired or derived from ground measurements) with appropriate classification strategies would be more effective in capturing LC dynamics and changes associated with the natural resources. Ancillary information would make the decision boundaries between the LC classes more widely separable, enabling classification with higher accuracy compared to conventional methods of RS data classification. In this work, we ascertain the possibility of improvement in classification accuracy of RS data with the addition of ancillary and derived geographical layers such as vegetation indices, temperature, digital elevation model, aspect, slope and texture, implemented in three different terrains of varying topography—urbanised landscape (Greater Bangalore), forested landscape (Western Ghats) and rugged terrain (Western Himalaya). The study showed that use of additional spatial ancillary and derived information significantly improved the classification accuracy compared to the classification of only original spectral bands. The analysis revealed that in a highly urbanised area with less vegetation cover and contrasting features, inclusion of elevation and texture increased the overall accuracy of IKONOS data classification to 88.72% (3.5% improvement), and inclusion of temperature, NDVI, EVI, elevation, slope, aspect, Panchromatic band along with texture measures, significantly increased the overall accuracy of Landsat ETM+ data classification to 83.15% (7.6% improvement). In a forested landscape with moderate elevation, temperature was useful in improving the overall accuracy by 6.7 to 88.26%, and in a rugged terrain with temperate climate, temperature, EVI, elevation, slope, aspect and Panchromatic band significantly improved the classification accuracy to 89.97% (10.84% improvement) compared to the classification of only original spectral bands, suggesting selection of appropriate ancillary data depending on the terrain.
KeywordsLand cover Classification Accuracy Ancillary layers DEM Vegetation indices Texture
We are grateful to (1) the NRDMS division, The Ministry of Science and Technology (DST), Government of India, (2) The Ministry of Environment, Forests and Climate Change, Government of India and (3) Indian Institute of Science for the financial and infrastructure support.
- Breiman, L., & Cutler, A. (2010). Breiman and Cutler’s random forests for classification and regression. Version 4.5-36, Repository—CRAN. http://ugrad.stat.ubc.ca/R/library/randomForest/html/00Index.html.
- Collins, W. J., Bellouin, N., Doutriaux-Boucher, M., Gedney, N., Halloran, P., Hinton, T., Hughes, J., Jones, C. D., Joshi, M., Liddicoat, S., Martin, G., O'Connor, F., Rae, J., Senior, C., Sitch, S., Totterdell, I., Wiltshire, A., & Woodward, S. (2011). Development and evaluation of an Earth-system model—HadGEM2. Geoscientific Model Development Discussion, 4, 997–1062.CrossRefGoogle Scholar
- Dópido, I., Villa, A., Plaza, A., & Gamba, P. (2011). A quantitative and comparative assessment of unmixing-based feature extraction techniques for hyperspectral image classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. doi: 10.1109/JSTARS.2011.2176721.Google Scholar
- GRASS GIS. (2017). r.texture, https://grass.osgeo.org/grass72/manuals/r.texture.html.
- Haack, B. N., Solomon, E. K., Bechdol, M. A., & Herold, N. D. (2002). Radar and optical data comparison/integration for urban delineation: A case study. Photogrammetric Engineering and Remote Sensing, 68, 1289–1296.Google Scholar
- Hung, M., & Ridd, M. K. (2002). A subpixel classifier for urban land-cover mapping based on a maximum-likelihood approach and expert system rules. Photogrammetric Engineering and Remote Sensing, 68, 1173–1180.Google Scholar
- Jensen, J. R., & Cowen, D. C. (1999). Remote sensing of urban/suburban infrastructure and socioeconomic attributes. Photogrammetric Engineering and Remote Sensing, 65, 611–622.Google Scholar
- Ramachandra, T. V., & Kumar, U. (2008). Wetlands of greater Bangalore, India: Automatic delineation through pattern classifiers. The Greendisk Environmental Journal, (International Electronic Journal), 1(26), 1–22.Google Scholar
- Ramachandra, T. V., & Kumar, U. (2009). Land surface temperature with land cover dynamics: Multi-resolution, spatio-temporal data analysis of Greater Bangalore. International Journal of Geoinformatics, 5(3), 43–53.Google Scholar
- Smith, M. O., Johnson, P. E., & Adams, J. B. (1985). Quantitative determination of mineral types and abundances from reflectance spectra using principal component analysis. In Proceedings of the Lunar and Planetary Science Conference, 90, (pp. 797–904).Google Scholar
- Turner, B. L., II, Meyer, W. B., & Skole, D. L. (1994). Global land-use/land-cover change: Towards and integrated study. Ambio, Integrating Earth System Science, 23(1), 91–95.Google Scholar