Classification and Extraction of Wetland Features Using Geoinformatics

Wetlands play significant role in maintaining the ecological balance of both abiotic & biotic life in inland and coastal environment. Wetlands occur where the water is at or near the surface of the land, or where the land is covered by the water. Wetlands have great importance more than one reason, because they charge aquifers, conserve moisture, act as pollution filter and are habitat for diversity. Hence, understanding of their occurrence, spatial extent of change in wetland environment is very important and can be monitored using satellite Remote Sensing technique. Accurately mapping wetland type and monitoring their dynamic changes provide the scientific foundation for wetland protection and restoration. The main aim of this study is to map and detect changes in wetland of study area using various digital classification techniques and compare accuracy of supervised & knowledge base classification. Spectral bands and spectral indices are used for extraction of wetland features. Indices are useful for delineating wetland features, and they are used in knowledge-based classification. A knowledge-based classification has been developed using decision tree approach for extraction of wetland features. Change detection analysis has been carried out to find change in wetland features. The study is carried out for pre monsoon and post monsoon period. It is observed that overall accuracy of knowledgebased classification is better than supervised classification.