Urbanization Analysis Using Spatial Support And Improved Random Forest Decision Tree Approach

Urbanization is the major strategy that leads to the loss of the various biodiversity and the homogenization of geological habitat. Thus, it is necessary to study and analyze the drastic changes occurred due to global urbanization periodically. The periodical assessment of urbanization gives rise to the development of various techniques and rules from several researchers. This paper is also a part of development over the urban-land cover analysis. It introduces a Budget in Random Forest Decision tree (RFDT) approach that preserves the statistical features and object boundaries and help in improving the classification accuracy. The system implements a spectral band segmentation method, which differentiates the remote sensory images as Land Cover (LC) and Land Use (LU). The proposed RF algorithm of decision trees attempts to provide an improved efficiency over the other existing methods that is obtained from the experimental verification of the earlier algorithms with the proposed. The comparison report shows the performance of the algorithm from 2007 to 2013 respectively.