Analysis of water quality parameters by hyperspectral imaging in Ganges River

Abstract

Optical dynamic properties of water have capacity to give a coherent outline of water quality, however its exactness relies on the samples gathered from water bodies. This work was performed to make utilization of remotely detected information for recognizing water quality parameters in view of optical dynamic properties of water. Airborne visible/infrared imaging spectrometer new generation (AVIRIS-NG) was guided over the Ganges River Buxer, Bihar, India gathering hyperspectral band of information. The image taken from nearly 10 km of the Ganges River with sample and line numbers 731 and 3190 respectively, having 425 spectral bands between 380 and 2510 nm wavelength with 5 nm sampling. Water test from 17 areas of Ganges River were gathered and dissected. By the utilization of ground truth information and combination of spectral band got from hyperspectral imaging, different spectral indices were readied which are valuable in evaluating chlorophyll-a, turbidity and aggregate phosphorus. The results show that the Pearson correlation between ground truth data and spectral ratio indices were stronger than the single spectral band. It is also clear that the various indices seem to associate with particular class of ground truth data. Spectral bands having wavelength 677, 702, 705, 671 and 742 nm are dominating the formation of spectral indices. The correlation and R2 for each investigated parameter were greater than 0.6 and 0.5 representing to develop a good linear model.