@book{6164,
	author = {Anju Francis (92316001)},
	title = {Change monitoring of agricultural lands using temporal hyperspectral datasets via feature extraction and classification techniques},
	note = {
Hyperspectral remote sensing, also known as imaging spectroscopy is one of the most significant 
breakthrough in remote sensing which Began in 1980’s. It is emerged as a technology for studying
earth surface materials by spectrally and spatially. In the field of agriculture hyperspectral remote
sensing offers study of species diversity, environmental stress, physiological features such as
photosynthetic activity, plant productivity, canopy biochemistry, biomass and plant transpiration.
Also for evaluation of vegetation stress, nutrient stress, moisture stress and crop growth models. In
this study, Hyperion image datasets covering the area of Kokrajhar, Assam in L1 R format. The
Hyperion images are preprocessed by the Hyperion tools provided as the extension to ENVI
software. Atmospheric corrections are performed by QUAC and FLAASH. Feature extraction was
performed to reduce the high dimensionality data into lower dimension via PCA transformation.
Classification was done using three techniques, Spectral Angle Mapper, Maximum Likelihood
Classification and Support Vector Machine based classification. In this three classification results,
the SVM classification after PCA transformation gave more accurate results. From the classified
results of SVM, the area for each land classes were estimated.
 

}
}