A 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. The authors propose a spectral reflectance curve-based approach to generate a linearized index. Using multispectral images, they first convert DN to reflectance, use KL transform and fuzzy logic to classify pixels, then determine a mapping function to generate a linearized index. An application of this approach to generate a linearized snow age index is shown. The linearized index can be further analyzed using time-varying signal analysis techniques to explore the spatial and temporal variations