Harnessing Spectral Libraries From AVIRIS-NG Data for Precise PFT Classification: A Deep Learning Approach.

Spectral libraries from AVIRIS-NG data enable 94% accuracy in plant functional type classification in Indian wildlife sanctuary.

• Why it matters: Accurate PFT classification is essential for monitoring forest health and biodiversity, but traditional methods lack spectral detail needed for precise differentiation.
• What they did: A spectral library from 130 plant species was created using AVIRIS-NG and ASD spectroradiometer data, with spectral features selected via clustering and JM analysis, then classified with three machine learning algorithms.
• The result: Gradient Boosted Machine achieved an overall accuracy of 0.94 and a Kappa of 0.93, demonstrating the potential of hyperspectral spectral libraries combined with advanced classifiers for precise PFT mapping.