What problem does it solve?
Working with hyperspectral and multispectral imagery in MATLAB involves dozens of specialized functions with subtle argument orders, object-type constraints, and format quirks that are easy to get wrong. This Skill gives an AI agent the exact syntax, constraints, and workflows of the Hyperspectral Imaging Library so it generates correct spectral imaging code on the first attempt.
Core Features & Use Cases
- Spectral I/O and Preparation: Read ENVI, NITF, GeoTIFF, Sentinel-2, Landsat, and ASTER data into hypercube or multicube objects, then crop, select bands, resample resolutions, and visualize with colorize.
- Calibration and Analysis: Run radiometric calibration, atmospheric correction, denoising, PCA/MNF reduction, endmember extraction, spectral unmixing, target and anomaly detection, spectral indices like NDVI, and segmentation.
- Labeling and Deep Learning: Use the Spectral Image Labeler app and groundTruthSpectralImage objects, then train 3-D CNNs, 1-D spectral CNNs, unmixing autoencoders, or DeepLabV3+ segmentation models.
- Use Case: Load a Sentinel-2 scene, resample all bands to 10 m, compute NDVI, and train a DeepLabV3+ model adapted for 12-channel input to segment land cover classes.
Quick Start
Ask the agent to load a hyperspectral ENVI file, compute an NDVI map, and detect anomalies using the spectral imaging skill.
