Urban Land Cover Analysis from Satellite Images

This is a study of modeling urban environments from TM satellite images. Urban environments are so heterogeneous. It is necessary to simplify them as combinations of basic land cover materials in order to enable quantitative studies. The V-I-S model proposed by Ridd in 1995 is a conceptual model to simplify urban environments as combination of three basic ground components: vegetation, impervious surface, and soil. Most urban grounds can be interpreted as combinations of these three basic components. This model is used by this study on Salt Lake City, Utah. As a step further, six ground components are selected as basic components of urban environments: two for vegetation, three for impervious surface, and one for soil. Percentages of these six ground components are extracted from a previously developed supervised classifier. Various charts and plots are generated to demonstrate the capacity of V-I-S composition on urban land cover analysis. Nine general urban features are selected and V-I-S composition extracted. Examinations of the V-I-S composition may reveal that most urban features have their own unique V-I-S composition, which is difficult to observe from satellite image or per-pixel classification methods. These unique V-I-s compositions may be a key factor for further urban land cover analysis.