In India, as per constitutional provisions, urban local bodies are mandated for administering, managing and preparing master / development plans. Mostly these plans are static maps with limited forecasting capabilities. Nevertheless there is a need for modeling the dynamics planning process to prevent ad-hoc decisions. In this context, regional models based on the information systems involving simulation for evolving location specific strategy and policy options are desirable.

From the middle of the last century, changes in the rural landscapes have been more sudden and have occurred at a broader scale as a result of the impacts of industrialisation, urbanization and, globalization in post nineties, which needs to be addressed at a local/micro scale (Antrop, 2005; Calvo-Iglesias et al., 2008).

The problem of sprawl needs to be addressed considering all disciplines with an integrative approach (TRB, 1998 and 2002; Gayda et al., 2003 and 2005). However, some attempts are made to capture sprawl in its spatial dimensions, which fail to capture sprawl process in other dimensions (like, travel times, pollution, resource usage, etc.) and also do not indicate their intensity (density metrics). Therefore for the better utilization of landscape and its features, regional planning need to account all classes of the landscape ranging from urban area to rural area, which is possible only when the data is available for all classes and on a temporal scale. Linear gradient analysis is, however, limited in capturing the spatial variation of land use patterns as it only examines patterns along a predefined direction (Yeh and Huang, 2009). The cities often results in non-linear morphologies, such as the concentric form (Jim and Chen, 2003; Tian et al., 2010), necessitating the analysis of spatial variation of land use patterns in concentric forms.

The indicators for achieving sustainable development have been evolved by Meadows (1998), and there isn’t yet any broad consensus on the appropriate indices representing all of the factors and disciplines. However, there is still a need for an improved understanding of urban change and its natural environmental and landscape consequences (Stephan and Friedrich, 2001; Stephan et al., 2005; Su et al., 2007a). It is imperative to address the change in landscape dynamics in various levels and through appropriate metrics or indicators for effective regional planning and sustainable utilization of natural resources.

CA has been used for simulating urban growth quite successfully mostly considering various driving forces that are responsible for sprawl. However some issues like the impact on ecology, energy, environment and economy for taking policy decisions have not been addressed effectively. In this context, the integration of agent-based models and CA models, where agent-based models would help in capturing the externalities driving the processes.

Development of new methods for retrieving information from mixed pixels based on Constrained Energy Minimisation, Mixture Tuned Matched Filtering, Adaptive Coherence Estimator, Spectral Feature Fitting, etc (Settle, 2002).

Development of new hard classification techniques based on (i) The object-oriented classification of remote sensing image takes the characteristics of the imaging spectrum and differences in geometric characteristics into account, which can extract more accurate image information. (ii) Genetic algorithms (GA) - GAs searches combination of multiple parameters in order to achieve the greatest level of satisfaction, either minimum or maximum, depending on the nature of the problem. It determines the knowledge rules for land-cover classification from remote sensing image datasets (Tseng et al., 2008).

Incorporation of endmember variability in the absence of endmembers in hyper-spectral data.

Development of Kernel-based methods for hyper-spectral image classification (Camps-valls and Bruzzone, 2005).

Feature extraction (such as roads, high rise buildings, etc.) using biologically inspired techniques - ant colonization, particle swarm optimisation, scale invariant feature transform (SIFT), etc. (Mikolajczyk and Schmid, 2005).