In situ estimation of water quality parameters in freshwater aquaculture ponds using hyperspectral imaging system

Estimation of chlorophyll-a concentration of different seasons in outdoor ponds using hyperspectral imaging

“…was obtained from the model of cross band ratio algorithm with six wavelengths than the other three algorithms because optimal wavelengths provided potential indication of water quality status in different The ability to predict Chl-a concentration by the model of cross band ratio algorithm with six wavelengths was better than those of two-band and three-band models (r 2 = 0.833, 078 and 0.862, 0.82, respectively), which were obtained by Abd-Elrahman et al [21] with a mobile, ground-based hyperspectral imaging sensor in working aquaculture ponds and by Cheng et al [22] with a similar field system in turbid lake water, respectively. In addition, the performance (r 2 = 0.88) for Chl-a estimation in turbid inland water by Song et al [49] was lower than that of the optimal wavelengths model in the current study.…”

“…(I) A two-band ratio approach in the form of R λ2 /R λ1 [21,45,46,47,48] , where λ1 is located around the phytoplankton absorption peak, and λ 2 is normally located between 700 and 720 nm [48] . Then, the R λ ratio can be calculated as:…”

“…(II) A three-band approach in the form of [21,47,49] , where CDOM and detrital absorptions at λ2 are close to those at λ1, and R λ3 is influenced by backscattering only, with minimal influence from water constituent absorption [48] . Thus, the R λ ratio can be calculated as:…”

“…With the development of HSI technology, there have been many studies about measuring the spectral reflectance of low Chl-a concentration in different inland water bodies using in situ 3 spectroradiometric measurements [20]. Abd-Elrahman et al determined Chl-a concentrations in working aquaculture ponds by a ground-based HSI sensor with coefficient of determination (r2) values of 0.833 and 0.862 for the two-band and three-band models, respectively [21]. Recently, Cheng et al estimated Chl-a concentration in pond water and Taihu lake in China using a similar field system by developing models with separating spectra based on Gaussian parameters with r2 values of 0.78 and root mean square errors (RMSE) of 4.80 mg/m3 [22].…”

Spectral reflectance is a reliable water-quality estimator for small, highly turbid wetlands

“…These remote sensing (RS) methods promise to be suitable for estimating WQPs in turbid waters over a large area several times a year; including chl-a (Sokoletsky et al 2011;Yacobi et al 2011;Shi et al 2013), prediction of phytoplankton species (Randolph et al 2008) and their biomass (Abd-Elrahman et al 2011), and/or TSS (Doxaran et al 2005Sterckx et al 2007;Ouillon et al 2008). Moreover, the spatial diversity within an individual water body can be evaluated and a time series can be obtained (Sudduth et al 2005).…”