Introduction
Welcome to H2M — hyperspectral research for pixel-level classification, without the steep learning curve.
Welcome to H2M. This documentation walks you through a research platform built to reduce hyperspectral data and support pixel-level classification: upload datasets, run classical machine learning models, and discover which spectral bands carry the most information for your task.
The goal is practical: researchers and students can use these workflows without heavy programming or deep ML prerequisites. For advanced use, H2M keeps data and results in one place, supports team collaboration, and records metrics for every saved training session so you can compare runs over time.
In HyperLabel, .mat is supported today; ENVI import is planned for a future release.
H2M is a research platform for reducing hyperspectral data and supporting pixel-level classification. Upload your datasets, run classical machine learning models, and surface the most informative spectral bands for your problem—without a steep programming or ML curve.
Three modules, one workflow
Pick a module to see what it does—then open its documentation when you are ready.
HyperIA
Upload datasets, train classical machine learning models, inspect performance metrics, and visualize the spectral bands that matter most for your signatures.
- Dataset upload and preprocessing
- Training sessions with saved metrics
- Band relevance and signature visualization