Publication

Abstract : The paper deals with Hyperspectral satellite data pre-processing and analysis for the application of finding the agricultural extent of a region. The Hyperion sensor of the National Aeronautics and Space Administration (NASA) which is carried by the Earth Observing 1 (EO-1) satellite, acts as the source of Hyperspectral images. These Hyperion images are of high spectral resolution with 242 bands and high spatial resolution of 30 meters. Before analyzing the images for agricultural applications or for any other applications, these images have to be pre-processed. Hyperion images come in different formats like level 0R (L0R), level 1R (L1R), level 1Gs (L1Gs) and level 1Gst (L1 Gst). For the purpose of this work the level 1R data of Kozhikode, Kerala is downloaded from the Earth explorer site provided by USGS. The Level 1R data, which is radiometrically corrected, has to undergo many pre-processing steps including geometric correction, destripping, subsetting of the images according to region of interest, noise and dimensionality reduction and finally atmospheric correction. ENVI 4.7 (the Environment for Visualizing Images) a powerful image processing system is used to perform the pre-processing of the data. ENVI provides an array of tools for processing hyperspectral data such as mapping tool, agricultural stress analysis tool, vegetation index tool, linear spectral unmixing and matched filtering. ENVI enables the provision to select GCPs (Ground Control Points) to georeference the image and the SPEAR (Spectral Processing Exploitation and Analysis Resource) tool in ENVI encompasses tools like Log residuals, Empirical line correction, Internal Average Relative Reflectance and many more for atmospheric correction and Vertical strip removal for destripping. Thus with pre-processing of available Hyperspectral data, the hyperspectral remote sensing which is rapidly advancing with the promise of land cover mapping, agricultural yield and many more applications can have more accuracy and greater predictability. © Research India Publications.