Urban land use mapping needs high resolution remotely sensed data. The pan-sharpened multi-spectral IKONOS imagery of 1 meter resolution is studied for urban land use classification of London, Ontario. With the increase of spatial resolution, between-class spectral confusion and within-class spectral variation increases. Spectral-based traditional image classification methods cannot be directly applied to the IKONOS data. In this study, a knowledge-based urban land use inferring method is proposed and tested on 36 image samples. These samples represent different land use categories, including residential, commercial, industrial, institutional, recreational (golf courses), forest and agricultural land use types. The proposed method includes two general steps. First, conventional multi-spectral classification method is applied to produce a preliminary land cover map. Then, urban land use information is inferred from land cover by a knowledge-based modelling process. The inferring rules involve the percentage composition ranges of compatible land cover categories for a certain land use class, the interrelationship of the compatible land cover, and excluding of un-compatible land covers. The result shows that the proposed method has successfully identified new and old residential, central commercial, institutional, recreational, forested and agricultural areas. However, it is difficult to separate suburban shopping malls and industrial areas in this study.
展开▼
