The conclusion drawn from the present study shows that there was an increase in bulk density, particle density, EC and soil available N and K while per cent pore space, water holding capacity, organic carbon while soil available P, Ca, Mg and S got declined. An increase in bacterial population and reduction in fungal and actinomycetes population was recorded. The fire has made a serious impact on vegetational composition due to the continuous promotion of retrogression by repeated fire occurrence. The effect of forest fire has heavily influenced the vegetation composition of the study site. The diversity indices and the vegetational study help to conclude that the type of fire occurring in the study site is crown fire and seriously damaging the tree strata in the vegetation. The analysis in the burnt and unburnt area showed that the forest has good regeneration potential especially after the monsoon period. However, the regeneration potential has been seriously affected due to regular occurrence of forest fire.The diversity indices and vegetational study helps to conclude that the regeneration potential of the study site is good enough to regain its original vegetation.

This book focuses on design and development of a unified framework for radiographic image analysis and interpretation with emphasis on defect classification. The framework is constructed by creating two ontologies namely process ontology and domain ontology. Process ontology structures information for performing gray scale image analysis and it is maintained as a knowledgebase. The images are analyzed by using the process plan which resulted in the enhanced images from which geometrical, statistical and textural features are extracted to construct domain ontology. In addition to that, details of welding defects in radiographic image are conceptualized and maintained in the domain ontology. The framework includes a flexible user interface through which the experts' knowledge of the domain can be uploaded into the knowledgebase. The knowledgebase search is minimized into single selection on the domain ontology using which the entire content of defects and type of defect appearing in the image can be visualized. The performance study shows that the developed unified framework outperformed the other two models viz. neural and statistical with respect to classification of defects.