Multispectral remote sensing innovations have been ordinarily utilized for acquiring and extracting information of Land Use Land (LULC) Cover features from earth’s surface in the past few decades. In a solitary perception, multispectral sensors acquire three to six spectral bands of data that ranges from the visible to Infra-Red (IR). This small window of spectral bands is an essential detriment to multispectral sensors and cannot be used in detailed LULCC. A new era started in remote sensing when hyperspectral sensors emerged as an excellent tool for gathering more than 200 spectral bands with narrow bandwidth that ranges from visible to Short Wave Infrared (SWIR) of the electromagnetic spectrum. Because of the fascinating detailed information available in hyperspectral data is spectrally over determined which is capable of distinguish spectrally similar material of earth surface features. Apart from advantage of hyperspectral data it has few limitations also. Fast computers, sensitive detectors and large data storage capacities are required which makes the acquisition and processing cumbersome and exorbitant Due to these limitations, very few number of space borne hyperspectral sensors are available till date while a lot of multispectral sensors are providing data with similar spatial resolution around the globe over the past few decades. Due to the availability of the vast multispectral datasets, it is indeed a requirement to simulate hyperspectral data utilizing multispectral data with a larger swath for detailed LULC studies. Simulated hyperspectral data will enable the identification and discrimination of subtle variations in the spectra of various features present over earth surface. In the present study spectral reconstruction approach is used for the simulation of hyperspectral data using EO-1 ALI multispectral data in open source programming environment. This technique was implemented in SCILAB open source software, which provides open customization environment for implementing application/theme/user specific algorithms. Atmospherically corrected multispectral and hyperspectral data is used in this study. EO-1 ALI Multispectral data along with ground spectra serves as an input for simulating hyperspectral data whereas atmospherically corrected EO-1 Hyperion data is used for validation. Over all 70 hyperspectral bands are simulated from EO-1 ALI multispectral bands and validated using EO-1 Hyperion bands by visual interpretation, classification, statistical approaches. The tone, texture, pattern, and shape of objects in both the images are appearing quite similar with a minimal difference. In the statistical approach most of the simulated spectral bands demonstrated very high correlation indicating better simulation of the hyperspectral bands. The last approach for validation is by performing classification on the simulated hyperspectral data and Hyperion data (with spectral bands of the wavelength range similar to simulated one) using the Spectral Angle Mapper method. The classified result of both the datasets are comparable and are able to distinguish major LULC classes. The study demonstrates potential of open source software’s for simulation of satellite datasets for detailed LULC classification.
Copyright © 2015. All Rights Reserved.