Contradiction Hidden in Values of Urban Raster Data

Abstract. This paper discusses that raster data identified by the independent largest fraction method may include both manifest and hidden contradictions, and describes the underlying mechanisms of contradictions. Using the examples of zones of land use and floor-area ratio, we demonstrate that the cells including more than three different zone category combinations might be identified with a non-existing combination of the zone categories. For the zones in which the ratio of contradictory cells is comparatively large, this problem can be quite significant. We model, therefore, the probability of contradictory identification using the adjacency relationship of different categories, and demonstrate the good fitness of the model using actual urban raster data. The results reveal that there can be a significant proportion of hidden contradictions.