Classification of remote sensing (RS) data accurately is a prerequisite for many environmental and socio-economic applications [1], such as urban change detection [2-3], urban heat islands [4-6], and estimation of biophysical, demographic, and socio-economic variables [7-8]. Satisfactory classification of RS data depends on many factors including (a) the characteristics of study area, (b) availability of suitable RS data, (c) ancillary and ground reference data, (d) proper use of variables and classification algorithms, (e) user’s experience with reference to the application and (f) time constraints [9]. Furthermore, diverse landscapes and terrain types have a mixture of both homogeneous and heterogeneous land cover(LC) classes and require supplemental environmental or geographical layers for improved classification accuracies. Increased spectral variation is common with high degree of spectral heterogeneity in complex landscapes [10]. For example, urban landscapes are composed of features having a complex mix of buildings, roads, flyovers, pavements, trees and lakes which are sometimes smaller than the medium spatial resolution sensors [11]. This creates mixed pixels, a common problem prevalent in residential areas where buildings, trees, lawns, concrete and asphalt all occur within a pixel, often responsible for low classification accuracy, and is a major challenge for selection of suitable image processing approaches over a large area. On the other hand, with the availability of fine spatial resolution data such as IKONOS Multispectral (MS) and Panchromatic (PAN) [12-13], numerous opportunities exist for urban studies. A major advantage is that they reduce the mixed pixel problem, providing the potential to extract much more detailed information of urban structures compared to mediumspatial resolution data.However, the most important issue associated with high spatial resolution data is that it is expensive and requires more time for data analysis than medium spatial resolution data [9]. Moreover, high spatial resolution often lead to high spectral variation within the same LC class, and the limited number of spectral bands, including the lack of a shortwave infrared band, leads to a high rate of spectral confusion resulting in poor classification performance [14]. In practice, data acquired from medium spatial resolution sensors such as Landsat TM/ETM+ or IRS LISS-III, being readily available for multiple dates, are commonly used for urban landscape analysis at a regional scale.

Reducing the spectral variation within the same LC and increasing the separability of different LC types are the keys for improving LC classification [10]. In this regard, different approaches such as sub-pixel classification [15-16], multi-sensor data integration [17], full spectral image classification [18-19], expert classification [20] have been used. Traditional per-pixel spectral-based supervised classification is based only on spectral signatures, but does not make use of rich spatial information inherent in the data [14]. Therefore, making full use of RS information along with ancillary information (acquired or derived environmental layers) would be an efficient way to improve classification accuracy.

The paper is organised as follows: Section 2 highlights the previous work done using additional geographical layers for improving classification accuracy, section 3 states the objectives, and section 4 describes the SMAP classification algorithm. Study area and Data are discussed in section 5, results from IKONOS and Landsat ETM+ are highlighted in section 6, followed by Discussion and Conclusion in section 7 and 8 respectively.