Mapping of Prosopis Juliflora by a Fusion assisted Pattern Based Classification

Abstract

This paper proposes a fusion assisted classification method to locate and map invasive plant species, Prosopis Juliflora using remote sensing techniques towards conservation of biodiversity. This is two stage method. At first, wavelet based fusion method is proposed for the multispectral image to produce high resolution multispectral image.In the second stage a texture based classification is performed ith pattern study.Minimum distance classifier is used to classify the input image based on weighted texture features. Accuracy is computed by collecting the ground truth points from the study site. Similar procedure is repeated for the Google Maps data and accuracy comparison of World View 2 and Google Maps is carried out. Thus World View 2 data outperform Google Maps data by achieving accuwracy of 80 percentage.