Minimum Noise Fraction (MNF) Based Mangrove Species Identification Using Hyperspectral Data (Hyperion) in Muthupet Lagoon, Tamil Nadu, South India

Abstract

The multispectral imagery contains three to ten bands and measures the reflected electromagnetic energy in the specific bands, whereas hyperspectral remote sensing provides 200 continuous spectral bands along with smaller bandwidth; hence, hyperspectral remote sensing techniques have great potential to detect a difference among the features of the earth. The Hyperion is a hyperspectral imager onboard Earth observing the EO-1 Platform launched by the National Aeronautics and Space Administration (NASA) in 2000. It consists of 242 spectral bands between 400 nm to 2500 nm in the electromagnetic spectrum. It is first attempted to identify the mangrove species in the Muthupet mangrove forest using a hyperspectral approach. The post level 1Gst of EO-1 (Hyperion) data has been exploited, radio-metrically corrected and geometrically re-sampled, and registered to the geographic map projection. Before analysing the hyperspectral data, proper organisational steps for pre-processing data have been carried out. The minimum noise fraction enabled data dimensionality to extract more information about the study area. The first ten MNF bands were selected to identify the unique pixels for various classes. The detailed field was made to collect ground truth points and built a spectral library for six prominent valid mangrove species such as Acanthus ilicifolius, Aegiceras corniculatum, Avecennia marina, Excoecaria agallocha, Rhizophora apiculata and Rhizhophora mucronata using the Analytic Spectral Device. The first two bands of minimum noise fraction effectively provided the pure pixels for Mangrove species. Post-processing can be computed for the classified image; the area of distribution of Aegiceras corniculatam is 1204.22 Ha, which is dominant occupancy. Areas of other species such as Rhizhophora apiculata, Avecennia marina, Excoecaria agallocha, Acanthus ilicifolius, Rhizhophora mucronata are 221.5 Ha, 861.3 Ha,24.8 Ha,114.9 Ha and 85.03 Ha respectively.