A double continuous approach to visualization and analysis of categorical maps |
| |
Authors: | Tomislav Hengl Corresponding author Dennis J J Walvoort Allan Brown? David G Rossiter |
| |
Institution: | 1. AGIS Centre, Faculty of Agriculture , Trg Sv. Trojstva 3 , 31000 Osijek, Croatia E-mail: hengl@pfos.hr hengl@itc.nl;3. ALTERRA , Building 101, Droevendaalsesteeg 3, 6708 PB Wageningen , Netherlands E-mail: dennis.walvoort@wur.nl;4. Department of Geo-information Processing , International Institute for Geo-information Science and Earth Observation (ITC) , PO Box 6, 7500 AA Enschede, Netherlands E-mail: brown@itc.nl;5. Department of Earth Systems Analysis , International Institute for Geo-information Science and Earth Observation (ITC) , PO Box 6, 7500 AA Enschede, Netherlands E-mail: rossiter@itc.nl |
| |
Abstract: | A method to visualize multiple membership maps, called ‘Colour mixture’ (CM) is described and compared with alternative techniques: defuzzification and Pixel mixture. Six landform parameters were used to derive the landform classes using supervised fuzzy k-means classification. The continuous categorical map is derived by GIS calculations with colours, where colour values are considered to represent the taxonomic space spanned by the attribute variables. Coordinates of the nine class centres (landform facets) were first transformed from multivariate to two-dimensional attribute space using factor analysis, and then projected on the Hue Saturation Intensity (HSI) colourwheel. The taxonomic value was coded with the Hue and confusion with Saturation. To improve visual impression, saturation was replaced with whiteness. Classes that were closer in attribute space were merged into similar generic colours. The CM technique limits the derived mixed-colour map to seven generic hues independently of the total number of classes, which provides a basis for automated generalization. The confusion index derived from the mixed-colour map was used to derive primary boundaries and to locate areas of higher taxonomic confusion. |
| |
Keywords: | Event structure perception Topology Spatio‐temporal reasoning Conceptual neighborhoods |
|
|