Land use-Land cover (LULC), Satellite images, Algorithm, Classification
Knowledge about land use and land cover (LU/LC) is necessary to plan, monitor and evaluate the developmental activities. Many methods of remote sensing technologies have been developed to detect LU/LC. Several digital classification approaches were used to classify the satellite dataset. Different methods have different principal algorithms and have different approaches. These technologies are required to be evaluated for better accuracy. In the current study, five techniques (isocluster unsupervised classification, maximum likelihood supervised classification, principal component based classification, spectral angle mapper based classification and decision tree approach based classification) have been used to identify LU/LC in Bhubaneswar city in India. The accuracy assessment is done on all the classification techniques to compare and choose the best suited algorithm. It was observed from the study that Decision Tree Classifier based classification showed best result with overall accuracy of 88.40 % and kappa value 0.855 thus suited well to the study area.
BYU ScholarsArchive Citation
"EVALUATION OF DIFFERENT TECHNIQUES TO DETECT LAND USE / LAND COVER OVER AN AREA,"
Journal of Spatial Hydrology: Vol. 14
, Article 2.
Available at: https://scholarsarchive.byu.edu/josh/vol14/iss2/2