PhotonLabeler

PhotonLabeler enables manual labeling and visual interpretation of ICESat-2 ATL03 geolocated photon data to produce labeled photon datasets for validation of ICESat-2 products (e.g., ATL08) and for algorithm training.

Key Features:

• Implementation: Developed in MATLAB.
• Input support: Reads ICESat-2 ATL03 Hierarchical Data Format (HDF) geolocated photon files.
• Visualization: Visualizes geolocated photon returns for inspection and interpretation.
• Manual classification: Provides point selection tools to manually classify individual photons into predefined target classes.
• Output formats: Saves labeled photon data in ASCII format.
• Session management: Supports saving and loading labeling sessions for persistence of labels.

Scientific Applications:

• Product validation: Generates ground truth labels for validating ICESat-2 mission products, particularly where airborne lidar reference data are lacking.
• Algorithm training and validation: Provides labeled datasets for training and validating algorithms that generate ICESat-2 data products.
• ATL08 evaluation: Applied in validation studies of the ATL08 land and vegetation product for canopy and terrain height estimates using night- and day-acquired ATL03 data (e.g., northwestern Zambia and Texas, USA).

Methodology:

Implemented in MATLAB; reads ATL03 HDF files; visualizes geolocated photons; enables manual point selection and classification of individual photons into predefined classes; saves labeled outputs to ASCII and allows saving/loading of labeling sessions.