Study of linearized index generation using multispectral images for environmental monitoring

Many indices have been used to study the environmental change. It is easy to monitor the change if an index provides a linear response to a physical or biophysical parameter. We proposed a spectral reflectance curves-based approach to generate a linearized index. Using multispectral images, we first converted DN to reflectance, used KL transform and fuzzy logic to classify pixels, then determined a mapping function to generate a linearized index. An application of this approach to generate a linearized snow age index is shown in this paper. The linearized index can be further analyzed using time-varying signal analysis techniques to explore the spatial and temporal variations.