Supervised Classification of Remote Sensed Data using Support Vector Machine
Support vector machines have been used as a classification method in various domains including and not restricted to species distribution and land cover detection. Support vector machines offer many key advantages like its capacity to handle huge feature spaces and its flexibility in selecting a similarity function. In this paper the support vector machine classification method is applied to remote sensed data. Two different formats of remote sensed data is considered for the same. The first format is a comma separated value format wherein a classification model is developed to predict whether a specific bird species belongs to Darjeeling area or any other region. The second format used is raster format which contains image of Andhra Pradesh state in India. Support vector machine classification method is used herein to classify the raster image into categories. One category represents land and the other water wherein green color is used to represent land and light blue color is used to represent water. Later the classifier is evaluated using kappa statistics and accuracy parameters.
