Natura

Last updated January 31, 2026

275

Innovation Areas

20,463

Inventors

Collaborations

Hyperspectral endmember extraction algorithms: NaturaRecent Research Landscape

Spectral mixing in hyperspectral data obscures sub-pixel material identification, leading to classification errors. This lever utilizes twin network architectures to decouple hybrid pixel components for precise material characterization.

What technical problems is Natura addressing in Hyperspectral endmember extraction algorithms?

Inaccurate land cover classification

(43)evidences

Spectral mixing and low spatial resolution prevent precise identification of specific terrain features like sandy beaches. Resolving this improves the reliability of automated remote sensing analysis.

Signal noise interference

(8)evidences

High background noise levels in spaceborne lidar data obscure actual surface returns. Eliminating these artifacts is essential for accurate bathymetric and topographic measurements.