The Impacts of Landscape Patterns on the Accuracy of Remotely Sensed Data Classification

The accuracy of the Land Use/Land Cover (LULC) data derived from remote sensing images is critical for many applications. Classification error is caused by the interaction of numerous factors, including landscape characteristics, sensor resolution, spectral overlap, preprocessing algorithms, and classification procedures. The purpose of this paper is to analyze the impacts of landscape characteristics on classification accuracy and to analyze the distribution of errors from a landscape pattern perspective. Logistic regression was employed to assess the impact of landscape characteristics on classification accuracy. Two landscape variables, patch size and heterogeneity, were calculated at the pixel’s level and sub-pixel’s level respectively and their effects were evaluated. The results indicate that classification accuracy increases as land cover patch size increases and as heterogeneity decreases. The effect of patch size is more important than heterogeneity and the impact of variables calculated at sub-pixel level is more important than pixel level.