Soft Computing Techniques Based Identification of Hotspots in Images

Use of soft computing techniques in disaster prediction and management has interested measures at different levels. Soft computing techniques have given the humans the power and capability to identify events preluding the occurrence of major disasters. Even though it¿s quite impossible to forecast all natural calamities, it is imperative to identify methods that can forewarn the possible occurrence of an event. Soft computing techniques have been used by researchers to provide these fore warning capabilities. This research work is such an attempt to extensively investigate and exploit the power of soft computing techniques to identify the location of volcano hot spots. In this proposed research work different soft computing techniques have been studied for their suitability in identifying volcano hot spots in satellite images. Multi spectral satellite data have been employed for image processing and analysis. Suitable modification and improvements have been suggested for existing techniques like KNN, SVM, ANN, to increase the prediction performance and accuracy. As a part of this research work an ANFIS based system classifier has also been developed and tested for its performance.