Vegetation abundance in and around Sultan Qaboos University (SQU), Muscat, Oman and its response in Geoeye-1 high resolution imagery are studied using different image processing methods namely band ratioing, parallelepiped, Normalized Difference Vegetation Index (NDVI), Principal Component Analysis (PCA), Minimum Noise Fraction (MNF) and linear spectral unmixing. The developed band ratioing using the near infrared band (B4) as numerator, divided by visible bands (B2 and B3) shows the vegetation are highly reflective in the near infrared and absorptive in the visible bands. The RGB images produced using PCA uncorrelated output bands with the band of NDVI and the MNF inherent dimensional bands with the band of NDVI determined the more representative components for vegetation based on the spectral variations. The PCA and MNF bands processed by linear spectral unmixing shows a plausible coherence with the field observations, proving that the processing strategy applied on multispectral data is useful for vegetation abundance analysis in the study area. The assessment of accuracy for the occurrence and spatial distribution of vegetation provided the overall accuracy of 100% with Kappa Coefficient = 1 in the matrix of parallelepiped and compared with the Spectral Angle Mapper (SAM) and Spectral Information Divergence (SID) algorithms. The evaluation of the results was done by visual comparison and accuracy assessment of the classified images and the field/ancillary data.

Vegetation abundance in and around Sultan Qaboos University (SQU), Muscat, Oman and its response in Geoeye-1 high resolution imagery are studied using different image processing methods namely band ratioing, parallelepiped, Normalized Difference Vegetation Index (NDVI), Principal Component Analysis ...