Performance Comparison of Band Ratio and Derivative Ratio Algorithms in Chlorophyll-A Estimation using Hyperspectral Data

Water body management is essential to provide quality water to the requirement of drinking, agriculture and industries. Estimation of water quality in dams, lakes, rivers is paramount activity in quality control. As phytoplankton concentration is based on the physical and chemical state of the water, it is a good indicator and its concentration represents water quality. As Chlorophyll -a(Chl-a) is a color pigment that exists in phytoplankton, estimation of Chl-a concentration provides the phytoplankton concentration. The conventional Chl-a measurement that includes sample collection Chl-a extraction to chemicals and testing at lab takes long time and has lengthy procedure. Measuring the reflectance by remote sensing methods provides Chl-a concentration in near real time. Many reflectance based algorithms like Band ratio, three band model, four band model and first order derivative based algorithms like derivative data and derivative ratio methods are developed to estimate Chl -a concentration. The reflectance of the Chl-a in red region is less as its absorption is more but it has no effect in Near Infrared region. The ratio of these two bands provides chlorophyll concentration. Similarly the derivative ratio in this region also reflects the Chl-a concentration. In this study, the performance of two band ratio methods of reflective and derivative were compared using high spectral resolution data acquired using in-situ spectroradiometer at Madiwala lake, Bangalore. Correlation study on these estimated values with measured values was conducted and Pearson correlation coefficient and determination coefficient were calculated. Relation between the reflectance, derivative values with chlorophyll-a concentration was established through regression study.