The scientific objectives of the mission are:
- Land cover classification
- Cloud-snow discrimination
- Biomass classification + wetlands determination
Land cover classification
Land cover classification relies on hyperspectral data’s ability to capture distinct spectral fingerprints of different materials. The data is divided into numerous spectral bands and statistical or machine learning algorithms are employed to identify patterns and associate them with specific land cover classes. For example, healthy vegetation shows strong absorption features in the visible and near-infrared regions, while water bodies have unique spectral responses. This detailed land cover classification is pivotal for land-use planning, environmental monitoring, and disaster management.
Cloud-snow discrimination
Hyperspectral imaging is capable of distinguishing between clouds and snow by examining spectral signatures. Clouds typically exhibit higher reflectance in visible and near- infrared wavelengths due to sunlight scattering by water droplets. Conversely, snow crystals display distinctive absorption features in the mid-infrared spectrum. This capability is instrumental in meteorology, aiding accurate weather forecasting and climate research.
Biomass classification
Hyperspectral imaging contributes significantly to biomass estimation and classification. Different plant species and vegetation types have unique spectral signatures, particularly in the visible and near-infrared ranges. Healthy vegetation have high reflectance in the near-infrared due to chlorophyll absorption, while stressed or unhealthy vegetation displays different spectral responses. By analyzing hyperspectral data and using machine learning techniques we can determine the composition and health of biomes.