Fusion Methods and Multi-classifiers to Improve Land Cover Estimation Using Remote Sensing Analysis

Abstract Adopting a low spatial resolution remote sensing imagery to get an accurate estimation of Land Use Cover is difficult task perform. Image fusion plays big role map the Cover. Therefore, This study aims find out refining method for estimating using these steps; (1) applying three pan-sharpening approaches combine panchromatic that has high with multispectral resolution, (2) employing five pixel-based classifier on and fused images; artificial neural net, support vector machine, parallelepiped, Mahalanobis distance spectral angle mapper, (3) make statistical comparison between image classification results. The Landsat-8 was adopted this research. There are twenty thematic maps were generated in study. A suitable reliable presented based most results validation performed by adopting confusion matrix method. made images all levels. It proved produced Gram–Schmidt Pan-sharpening classified machine result among other classifiers, it overall accuracy about (99.85%) kappa coefficient (0.98). However, mapper algorithm lowest compared methods, 53.41% 0.48. proposed procedure useful industry academic side purposes. In addition, also good tool analysts researchers, who could interest extend technique employ different datasets regions.