Methods and Algorithms
Information Extraction regarding Environmental change using Remote Sensing imagery was achieved by developing the following methodologies:
- Wavelet Techniques for reducing unwanted multiplication effects on signals obtained from hyper and multi spectral sensors.
- Advanced Unmixing methods utilizing both hyper/ multi spectral and multi-temporal data allowing estimation of cover fractions in heterogeneous environments.
- Multi-temporal backscattering index for deriving Soil moisture from SAR data.
- Phenological Subtraction technique for deriving Life-form compositions.
- Reconstruction techniques allowing derivation of information regarding past vegetation and soil patterns from single date photographs.
- Multi-scale and multi-temporal analysis of soil drying rates for extracting information regarding soil groups.
- Water cloud model facilitating integration of different sensory data for estimating Areal Above Ground Herbaceous biomass.
- Knowledge based Expert systems for integrating sensory and existing thematic information for mapping land cover / land use.
- Mapping shrublands biomass using spectral data, rainfall information and Water Use efficiency information.
- Mapping and analysis of long-term trends in Water Use Efficiency along Climatic Gradients.
Information extraction regarding spatial and temporal pattern was achieved by developing the following techniques:
- Landscape change analysis for revealing rates of land use and land cover transformation and their impact on open / semi-natural landscapes.
- Cellular Automata simulation of Patch Pattern dynamics allowing insights into modes of vegetation spread strategies on these patterns.
- Combining Monte Carlo and FARSITE techniques for simulating fire spread as the main source of anthropogenic disturbance .
- Autocorrelation techniques for pattern analysis and for revealing core areas of change.