Please use this identifier to cite or link to this item: http://nopr.niscair.res.in/handle/123456789/55639
Title: Efficient pre-processing and retrieval of reflectance using calibration modules for Hyperspectral satellite data (Hyperion) and denoising of Hyperspectral reference spectra using wavelet based Adaptive Bilateral Filtering (HABF) -A case study on mangroves forest in Muthupet lagoon, Tamil Nadu
Authors: Selvaraj, A
Saravanan, S
Keywords: Bilateral filter;Destriping;Digital number to radiance;FLAASH;Hyperion data
Issue Date: Oct-2020
Publisher: NISCAIR-CSIR, India
Abstract: The article invoves in the preprocessing of hyperion data, and denoising of hyperspectral reference spectral library, which is generated from the spectro-radiometer. The atmospheric correction module like Fast Line-of-sight of Atmospheric Analysis of Spectral Hypercubes provided the smooth absolute spectral profile. In connection to that, the Hybrid Adaptive Bilateral Filter (HABF) is proposed for denoising of field based spectral library, and laboratory based spectral library. To implement the proposed algorithm, spectral library of mangrove has been exploited, which is collected from Muthupet mangrove forest, and spectra of particular species is generated from Field-spectro-radiometer with wavelength ranges from 350 nm to 2500 nm, and 10 nm band width. This spectral library has been used as an input signal since it contains noise across wavelength ranges from 350 – 450 nm, 1000 – 1200 nm, and 2000 – 2500 nm due to atmospheric conditions. These noises can be removed effectively by the proposed wavelet based HABF techniques and conventional method of denoising. The performance of each method was compared with performance evaluation parameter such as PSNR and MSE
Page(s): 1619-1626
URI: http://nopr.niscair.res.in/handle/123456789/55639
ISSN: 2582-6727 (Online); 2582-6506 (Print)
Appears in Collections:IJMS Vol.49(10) [October 2020]

Files in This Item:
File Description SizeFormat 
IJMS 49(10) 1619-1626.pdf531.79 kBAdobe PDFView/Open


Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.