Development of A Model for Detection of Saline Blanks Amongst Mangrove Species on Hyperspectral Image Data

In this study we apply hyperspectral imagery to identify saline blank patterns within the mixed mangrove forest of Sunderban Bio-geographic Province, West Bengal, India. We use derivative analysis to identify hyperspectral wavelengths that are sensitive to the presence of minerals comprising saline blanks. These wavelengths have been considered for development of a novel saline blank identification model. The wavelength showing derivative value with maximum absorption in the SWIR region at 1780 nm and maximum reflection in the red region at 690 nm has been extracted for development of saline blank index. This index has been compared with the existing salinity indices and a detailed analysis has been carried out. It is found that the index outperforms existing salinity indices – normalized differential salinity index and salinity index, and accurately detects the saline blank areas of Henry Island of the Sunderbans Delta. The accuracy of pixels identified as saline blank pixels has been assessed by comparing the overall accuracy with other existing indices. Physical sampling has also been carried out and the salinity results have been compared with the image-derived results.