Object identification and classification in a high resolution satellite data using data mining techniques for knowledge extraction

A mark improvement in spatial technology has happened over the last decade. Classification of objects, detection of change and interpretation of information from high resolution satellite imagery can only be carried out by human interpretation and using specialized tools capable of evaluating pixel level details. In this thesis, spatial information is stored using data mining. Automated classification and identification of objects are archived based on historical data set. Any object can be defined by three primary element color, shape and size. Secondary parameters such as texture, shadow and association help in increasing accuracy. Data mining task govern the object identification process. A Model based approach along with clustering algorithms such as K-means, Expectation maximization and Scalable clustering analyze and identify objects from the huge volumes of historical and newly acquired satellite data-sets. This would help the end user to gather and analyze relevant information rapidly and accurately