Ch.Mallikarjuna Rao
Gokaraju Rangaraju Institute of Engineering and Technology , India ,
Title: Segmentation Based Clustering for Knowledge Discovery from Spatial Data
Biography
Biography: Ch.Mallikarjuna Rao
Abstract
Global Positioning Systems (GPS) and other data acquisition systems have been collecting huge amount of geographical data that is growing in exponential growth. Mining such data can extract unknown and latent information from spatial datasets that are characterized by complexity, dimensionality and large size. However, it is challenging to do so. Towards this end geographical knowledge discovery through spatial data mining has emerged as an attractive field that provides methods to leverage useful applications. Remote sensing imagery is the rich source of geographical data. Analyzing such data can provide actionable knowledge for making strategic decisions. In this paper we proposed a methodology that is used to perform clustering on remote sensing images. The data sets are collected through World Wind application provided by NASA. The images are with .TIF extension. The methodology includes feature extraction, training, building classifier and cluster analysis. We built a prototype application that demonstrates the proof of concept. The implementation has taken native method support from Fiji and Weka to realize the proposed methodology. The empirical results revealed that the spatial clustering is made with high accuracy.